• TABLE OF CONTENTS
HIDE
 Front Cover
 Title Page
 Acknowledgement
 Table of Contents
 List of Tables
 List of Figures
 Highlights
 Introduction
 Perspectives and methodology
 Presentation analysis
 Implications of the findings and...
 Reference
 Appendix
 Back Cover






Title: Study of juvenile offenders in St. Thomas and St. Croix, USVI
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/CA01300584/00001
 Material Information
Title: Study of juvenile offenders in St. Thomas and St. Croix, USVI
Physical Description: Book
Language: English
Creator: University of the Virgin Islands. Eastern Caribbean Center
Virgin Islands of the United States. Law Enforcement Planning Commission ( Contributor )
Affiliation: University of the Virgin Islands -- Eastern Caribbean Center
Publisher: Eastern Caribbean Center, University of the Virgin Islands
Publication Date: 2003
 Subjects
Subject: Caribbean   ( lcsh )
Spatial Coverage: North America -- United States Virgin Islands
Caribbean
 Record Information
Bibliographic ID: CA01300584
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.

Downloads

This item has the following downloads:

SJO ( PDF )


Table of Contents
    Front Cover
        Front Cover
    Title Page
        Title Page
    Acknowledgement
        Page iii
        Page iv
    Table of Contents
        Page v
        Page vi
    List of Tables
        Page vii
        Page viii
        Page ix
    List of Figures
        Page x
    Highlights
        Page xi
        Page xii
        Page xiii
        Page xiv
        Page xv
        Page xvi
        Page xvii
        Page xviii
    Introduction
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
    Perspectives and methodology
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
    Presentation analysis
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
    Implications of the findings and recommendations
        Page 89
        Page 90
    Reference
        Page 91
    Appendix
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
    Back Cover
        Page 99
Full Text





A STUDY OF JUVENILE OFFENDERS
in St Thomas and St Croix, USVI















A STUDY OF JUVENILE OFFENDERS
in St Thomas and St Croix, USVI





Submitted to:
The Law Enforcement Planning Commission
Government of the US Virgin Islands





Prepared by:

Frank L. Mills, PhD
Director
Annette Gumbs Marcy Mason, PhD
Research A analyst III Research A sst. Professor of Statistics
James Richardson
Research Analyst I

Eastern Caribbean Center
University of the Virgin Islands


June 2003










Acknowledgements


The 1990s was a particularly stressful
decade for Virgin Islands parents,
school administrators and teachers.
Reports in the media of the increasing
level and frequency of violence among
youth across the Territory shocked
adult residents into the realization that
the culture of juveniles was very
different from that within which they
were raised. Perhaps the low points
were reached when it became evident
that dangerous weapons, including
firearms, were being smuggled by high
school students on to the campuses,
and when this deviant behavior
ultimately led to homicide.

Inevitably, many in the community
looked to our elected leaders for
answers to the questions concerning
the estrangement of our youth. It was
with this background that former Virgin
Islands Senator Allie "Allison"
Petrus-who campaigned on a
platform of youth issues-invited the
Director of the Eastern Caribbean
Center (ECC) to his office to discuss
the possibility of conducting a
penetrating study of juvenile
behaviors. It was not until funds to
support this activity were identified by
Ms Helene Smollett of the Law
Enforcement Planning Commission
(LEPC) that this study got underway.

The first formal meeting to launch this
effort was held at the LEPC offices in
Sub-Base, St Thomas under the
patronage of former Commissioner of
Police Ramon Davila. Also present
at that meeting was former
Commissioner of the Department of


Human Services Ms Catherine Mills.
She was very instrumental in
suggesting that analysis should be
undertaken to flesh out an explanation
why some youth become recidivist and
others do not.

When the decision was made to have
data collected on youthful offenders
from records within the police
department, two officers stand out for
their contribution in supervising this
task. They are Sgt Merlin Christian
of the St Thomas Juvenile
Investigation Bureau, and
Sgt Thomas Hannah of the St Croix
Youth Investigation Bureau. ECC also
acknowledges the endeavors of the
many other officers on both islands
who assisted with this aspect of the
project.

There were many juveniles for whom
records existed in both the Police
Department and the Department of
Human Services, and it was
considered critical to link the two.
ECC had the unhesitating support of
Human Services Commissioner
Sedonie Halbert, and it was no less
fortunate in having the enthusiastic
encouragement and willingness to
help of Ms Janet Turnbull-Krigger in
the St Thomas office. Ms Carol
Battuelo in the St Croix office was
equally supportive as she supervised
the collection of data.

This report has benefited enormously
from the input of several staff
members of the Eastern Caribbean
Center during the time it required to









complete it. One of our former staff,
Ms Carmen Rogers-Green, provided
significant input in the early stages of
the project, including survey
instrument development, data
collection, tabulation and the initial
writing of the report. Mr Ramesh
Srivastava's statistical expertise
proved invaluable in proofing and
certifying many of the statistical tables
that appear in the report. And even
though Ms Marsha Penn's tenure with
the Center was not very long, her
willing contribution in helping to bring
the report to near completion is
sincerely appreciated.

I particularly wish to recognize the
current members of staff who have
endured my exhortations to get this
report into the hands of those who
desire to understand what the data on
juvenile offenses are saying about our
community. Long-term staff member
Ms Annette Gumbs deserves
applause for the valuable assistance
she provided through all phases of
project planning, data collection and
processing, as well as in the


development of the report. Our
relatively new staff member Mr James
Richardson is also complimented for
introducing new energy into the
project, especially through his skills in
producing our first single integrated
electronic report. I value no less the
enthusiasm and assiduousness of our
latest staff member, Dr Marcy Mason,
who not only reviewed the entire draft
report, but who contributed through
her skills several improvements in
formatting, presentation and
interpretation.

Finally, ECC could not have completed
its work without the support of several
institutional offices, including: the
office of Vice Provost Dr Henry
Smith, that of Financial Management,
Human Resources and Accounting.

I thank all of those mentioned above,
as well as those I may have
inadvertently overlooked, and hope
that through this work we will all help
to improve the quality of life for the
youth in the Virgin Islands community.


Frank L. Mills
Director












TABLE OF CONTENTS


A C K N O W LE D G E M E N T S ............................................................................................. iii

L IST O F TA B LES ............... .................................................................. ...... ........... vii

L IS T O F F IG U R E S .................. ...................................... ........ ........ ........ ........ x

HIGHLIGHTS .................................... ..................... xi

I. INTRODUCTION
1.1 Background to the Study..................................................... .... .............. 1
1.1.1 H historical and Social Change ........................................... ........... 1
1.1.1.1 Som e Effects of Im m igration ............................. ....................... 1
1.1.2 Risk Factors Affecting Juveniles ........................................ .......... 4
1.1.3 At-Risk Indicators in the US Virgin Islands ........................................ 5
1.1.4 Y outh Crim e Trends......................................................... ........ 5
1.1.5 US Virgin Islands Youth Rehabilitation Center .................................... 7
1.1.6 Departm ent of Human Services ........................................ .......... 9
1.1.7 Juvenile Units ................ ............. .......... .......... .......... .... ............. 10
1.2 Purpose of the L E P C Study ............................................................................... 12
1.3 Scope of the Study ...................... .............. .......... .. 13
1.4 Survey Goals ........................................ 14
II. PERSPECTIVES AND METHODOLOGY
2.1 Delinquency and Antisocial Behavior ....................................... ..... ........ 15
2.1.1 Causes and R ecom m endations ................................... .................... ... 15
2.1.2 Development, Persistence and Pathways of Early Problem Behaviors .. 15
2.1.3 Explanation of the Causes of Antisocial Behavior ............................. 17
2.1.4 Prenatal and Perinatal Factors .............. ............................ ......... 17
2.1.5 H yperactivity .......................................................................... ..... 17
2.1.6 Intelligence and Attainm ent ........................... ..................................... 18
2.1.7 Parental Supervision, Discipline, and Attitude ...................................... 18
2.1.8 B roken H om es.................... .. .... .................... .. ...... ..... ............. 19
2.1.9 Large Fam ily Size .............. ............... ........... ........ ............. 19
2.2 M ethodology .................. ................................... ............ ..... ... 20
2.2.1 Steps in Developing the Framework of the Study .............................. 20
2.2.2 Data Collection Instrument .............. ...... ...................... ........... 21
2.2.3 Selection of the Population ............................................. ... ................ 21
2.2.4 Data Collection Procedures............... ...................... 21
2.2.5 Data Availability and Deficiencies ................................. ............. 22
2.2.6 D ata A analysis Procedures ........................................ ....... ............. 22
2.2.7 O organization of the R eport.............................................. .. ................. 22
2.2.8 L im stations of the Study.................................................. ... ................ 22









III. PRESENTATION ANALYSIS
St Thomas Data: Description and Analysis .............. ............................ 25
3.1 Juvenile Investigation Bureau Data Variables.................... ............... 25
3.1.1 St Thomas Data Description: Juvenile Investigation Bureau ............. 25
3.1.2 St Thomas: Juvenile Investigation Bureau Data Relationships .............. 27
3.1.3 Predicting Recidivism with the St Thomas Juvenile Investigation
B ureau D ata.............................. ............. 33
1.1.3.1 A approach ........................................ 33
2.1.3.2 Methodology .............................................................. 34
3.1.3.3 Definition of Variables .............. ...... ...................... .......... 34
4.1.3.4 The M odel .................. ... ........ .............. .......... ..... ........ 35
5.1.3.5 Estimates of the St Thomas Juvenile Investigation Bureau Data... 35
6.1.3.6 The Prediction of Recidivism ........................................... ...... 37
3.2 Department of Human Services Data Variables ................................... 40
3.2.1 St Thomas Data Description: Department of Human Services ............. 41
3.2.2 St Thomas: Department of Human Services Data Relationships ............ 45
3.2.3 Predicting Recidivism with the St Thomas Department of Human
S erv ice s D ata ..................... .... .......... ... ......... ........................... 4 8
1.2.3.1 Estimates of the St Thomas Department of Human Services
D ata ............................................ .. ................... 4 8
3.2.3.2 The Prediction of Recidivism ............. ......................... .............. 52
St Croix: Data Description and Analysis ....................................................... 53
3.3 Youth Investigation Bureau Data Variables ......................... ............. 53
3.3.1 St Croix Data Description: Youth Investigation Bureau......................... 53
3.3.2 St Croix: Youth Investigation Bureau Data Relationships .................. 55
3.3.3 Predicting Recidivism with the St Croix Youth Investigation Bureau
D ata ................... ... ..... ...... .................... ... .......... 63
3.3.3.1 Definition of the Variables................... .. ..................... 64
3.3.3.2 Estimates of the St Croix Youth Investigation Bureau Data ........ 64
3.3.3.3 The Prediction of Recidivism .............. .............................. ....... 65
3.4 Department of Human Services Data Variables .............................. ................ 68
3.4.1 St Croix Data Description: Department of Human Services .............. 68
3.4.2 St Croix: Department of Human Services Data Relationships .............. 72
3.4.3 Predicting Recidivism with the St Croix Department of Human
S erv ices D ata ........................... ........... ......... .......... ........... .... 80
1.4.3.1 Estimates of the St Croix Department of Human Services
D ata ................... .............. .......................... .... ........ 80
3.4.3.2 The Prediction of R ecidivism ........................................................ 80
3.5 Sum m ary and Overview for the Territory.................................. ................... 84
3.5.1 Juvenile U nits ............................................ ... .. .. .... .. ........ .. 84
3.5.2 Department of Human Services .... .......... ..................... ............... 85
IV. IMPLICATIONS OF THE FINDINGS AND RECOMMENDATIONS ................ 89

R EFER EN CE S ................ .... .................................. .............. .. ........ .... ............. 91

APPENDIX .......................................................... 92









LIST OF TABLES


Table 1.1
Table 1.2
Table 1.3
Table 1.4
Table 1.5
Table 3.1.1
Table 3.1.2
Table 3.1.3
Table 3.1.4
Table 3.1.5
Table 3.1.6
Table 3.1.7
Table 3.1.8
Table 3.1.9
Table 3.1.10
Table 3.1.11
Table 3.1.12
Table 3.1.13
Table 3.1.14
Table 3.1.15
Table 3.1.16
Table 3.1.17
Table 3.1.18
Table 3.1.19
Table 3.1.3.1

Table 3.1.3.2

Table 3.2.1
Table 3.2.2
Table 3.2.3
Table 3.2.4
Table 3.2.5
Table 3.2.6
Table 3.2.7
Table 3.2.8
Table 3.2.9
Table 3.2.10
Table 3.2.11
Table 3.2.12
Table 3.2.13
Table 3.2.14


Place of Birth by Year: 1980, 1990, 1995, 1997 ....................... ................ ......... 3
Possible Indicators of Juvenile Delinquent Behavior ............................................ 5
Three Population Types Served by the Youth Rehabilitation Center: FY 1999 ..... 8
Characteristics of Youths Remanded at YRC FY 1999 ....................................... 8
Juveniles Remanded Outside YRC FY 1999.... ................................... 9
Y ear of First Offense ..................... ... ............. ........ .......................... ............... 25
Type of First Offense ..................... ... ............... ...... .......................... ............... 26
W eapon U sed in F irst O offense ............................................................................... 26
Total N um ber of O offenses ........................................... ........................... ... 26
Total Number of Offenses by Place of Birth ............................................. 27
Type of First Offense by Place of Birth ............................... ........... 27
Type of Weapon Used by Place of Birth ........... ....... .................... 28
Duration Between First and Second Offense .............. .......................... ...... ...... 28
Age at First Offense by Sex ...................................................... ........... 29
Place of B irth of Juvenile by Sex ........................................ .... ..... ............. 29
Age at First Offense by Juvenile's Place of Birth .............................................. 30
Age at First Offense by Mother's Place of Birth ......................... ............. 30
Age at First Offense by Father's Place of Birth ............................... .... ............ 31
Juv enile's R ace by Sex .... .................................................................. .... 3 1
Juvenile's Guardian by Juvenile's Sex .................................................. 32
School E nrollm ent ..................... .... ............................ ..................... ...................... 32
School Type ........................................ 32
M other's Place of Birth by Sex of Juvenile ............. ............ ....... .............. 33
M other's Em ploym ent by Sex of Juvenile ........................................ ............. ... 33
Parameter Estimates in Logistic Regression Predicting Recidivism:
St Thom as, Juvenile Investigation Bureau ........................................ ............. ... 36
Predicted Probabilities in Logistic Regression: St Thomas, Juvenile
Investigation Bureau ............... ...... ................................... .......... 38
Juvenile's Sex .......................................................... 41
Im m migration Status ............... ..................... .................... .. ........ ..... ............ 41
Age of Juvenile ..................... ........ .................... ....... ............ 41
Juvenile's Place of Birth ............................ ............................................................ 42
Ju v en ile's R ace .......................................................................... 4 2
H household Status of Parents .............................................. ............ ............ 42
A ge of Parents ............................................ ............ 43
Place of Birth of Parents ............................ ............................................................ 43
M arital Status of Parents ................................................. ........................... 43
E m ploym ent Status of P parents ............................................................................... 43
Juvenile's G uardian ........ ............. .... ...................... .... ................ 44
Marital Relationship ofNatural Parents .................................................. 44
Siblings in the H household ............................................... ........................... 44
School Enrollment .. ........................................... .......... .. 44









Table 3.2.15
Table 3.2.16
Table 3.2.17
Table 3.2.18
Table 3.2.19
Table 3.2.20
Table 3.2.21
Table 3.2.22
Table 3.2.23
Table 3.2.24
Table 3.2.25
Table 3.2.26
Table 3.2.3.1

Table 3.2.3.2

Table 3.3.1
Table 3.3.2
Table 3.3.3
Table 3.3.4
Table 3.3.5
Table 3.3.6
Table 3.3.7
Table 3.3.8
Table 3.3.9
Table 3.3.10
Table 3.3.11
Table 3.3.12
Table 3.3.13
Table 3.3.14
Table 3.3.15
Table 3.3.16
Table 3.3.17
Table 3.3.18
Table 3.3.19
Table 3.3.20
Table 3.3.3.1

Table 3.3.3.2

Table 3.4.1
Table 3.4.2
Table 3.4.3
Table 3.4.4
Table 3.4.5
Table 3.4.6


Current or Highest Grade Completed ..............................................................
Juvenile's E m ploym ent Status ............................. ............................................
C u rren t Job .................................................
N um ber of Children in the H household .............................................. ...............
Juvenile's C hildren...........................................
Juvenile R responsible for Childcare ......................... ........................................
Juvenile's Psychological Assessment.............................................................
Juvenile's Sexual A activity ................................ ...............................................
Juvenile F orced to have Sex ............................... .............................................
Juv enile's Sex P artn ers ................................... .. .................. .......... ...............
P re-O u tcom e Statu s ....................................... .. .................. .......... .................
O utcom e Statu s ..............................................
Parameter Estimates in Logistic Regression Predicting Recidivism:
St Thomas, Department of Human Services Data ...........................................
Predicted Probabilities in Logistic Regression: St Thomas, Department
of H um an Services D ata .................................. ...............................................
Y ear of F irst O offense ......................................... .. .. .. .......... ...............
Type of First O offense .......... ........... .............. ............... .. ..................
W eapon U sed in First O offense ............................. ............................................
T otal N um ber of O offenses ................................. ..............................................
A ge at F irst O offense by Sex ................................ .............................................
Juv enile's P lace of B irth .................................. ................................................
Juvenile's P lace of B irth by Sex ............................................................................
Age at First Offense by Juvenile's Place of Birth .................................................
Age at First Offense by Mother's Place of Birth ...................................................
Age at First Offense by Father's Place of Birth .....................................................
Ju v en ile 's R ace b y S ex ........................................................................................
Juvenile's G uardian by Sex ............................... ..............................................
S ch ool E nrollm ent ........................................... .. .. .. .......... ................
S ch o ol T y p e ................................................
Mother's Place of Birth by Sex of Juvenile .....................................................
Mother's Employment by Sex of Juvenile ......................................................
Father's Place of Birth by Sex of Juvenile ......................................................
Father's Employment by Sex of Juvenile ........................................................
Recidivism and Relatives Charged with Offenses .........................................
Juvenile's Age by Current or Highest Grade ...................................................
Parameter Estimates in Logistic Regression Predicting Recidivism: St Croix,
Youth Investigation Bureau Data ......................................... ...............
Predicted Probabilities in Logistic Regression: St Croix, Youth Investigation
Bureau Data ............ ............................ ........ ............
Ju v en ile's S ex ...............................................
Im m ig ration Statu s ...................... .... .................................. .. ........ .................
A ge of Juvenile ..............................................
Juv enile's P lace of B irth .................................. ................................................
Ju v en ile's R ace ....................... ......................................................
Household Status of Parents .....................................









Table 3.4.7
Table 3.4.8
Table 3.4.9
Table 3.4.10
Table 3.4.11
Table 3.4.12
Table 3.4.13
Table 3.4.14
Table 3.4.15
Table 3.4.16
Table 3.4.17
Table 3.4.18
Table 3.4.19
Table 3.4.20
Table 3.4.21
Table 3.4.22
Table 3.4.23
Table 3.4.24
Table 3.4.25
Table 3.4.26
Table 3.4.27
Table 3.4.28
Table 3.4.3.1

Table 3.4.3.2

Table Al
Table A2
Table A3
Table A4
Table A5


A ge of the P parents ................................ .... .... ......... .......................... 70
M arital Status of Parents ................................................. ........................... 70
Juvenile's Guardian ............................ ................ ............ 71
Marital Relationship of Natural Parents .............................................. ............ 71
Siblings in the H household ............................................... ........................... 71
School E nrollm ent ............................. ........................ ...................... ..... ........ .. 72
Current or Highest Grade Completed ............................... ....................... 72
Juvenile's Age by Current or Highest Grade ............................................. 73
Juvenile's Em ploym ent Status ............................................ ....................... 74
Current Job ................................................ 74
Children in the H household ........................... ......................................................... 74
Juvenile's Children ..... ............................................................................ 74
Juvenile R responsible for Childcare ........................................ ....... ............ 74
Juvenile's Psychological A ssessm ent ................................ ....................... 75
Juvenile's Sexual A activity ........................... ......................................................... 75
Forced Sex ........................................................... 75
Juvenile's Sex Partners ............................ .............................................................. 76
Pre-Outcom e Status ..................... ... ....................... .......................... .. .... ...... 76
Type of O offense ...................... ... ........................... ................................ ............ 77
Y ear of O ffen se ........................................................................ 7 8
Outcom e of Offense ..................... ... ......................... ............................ .... ...... 79
B biological Father K now n ............................................................ ........................ 80
Parameter Estimates in Logistic Regression Predicting Recidivism: St Croix,
Departm ent of H um an Services Data.................................................................. 81
Predicted Probabilities in Logistic Regression: St Croix, Department of
H um an Services Data ............... .... .............................. ... ......... ........ ...... 82
Parent's Place of Birth St Thomas............. ..................... ................ 92
Type of Offense St Thomas.............................................. ......... 92
Type of Weapon St Thomas ................ ....................... ............. 93
Year of Offense St Thomas............................................ ............ 94
Outcom e of O offense St Croix.......... ................. .............. .............. 96









LIST OF FIGURES


Figure 3.1.3.1
Figure 3.1.3.2
Figure 3.2.3.1
Figure 3.3.3.1
Figure 3.3.3.2
Figure 3.4.3.1


Predicted Chances of Recidivism: St Thomas, JIB ....................................... 39
First-time Arrests of Juveniles in St Thomas: JIB............................................ 40
Predicted Chances of Recidivism: St Thomas, DHS ...................................... 52
Predicted Chances of Recidivism: St Croix, YIB ........................................ 67
Age at First Offense: St Croix YIB Data......... .. .................. .................... 67
Predicted Chances of Recidivism: St Croix, DHS ......................................... 83









HIGHLIGHTS


ST THOMAS

Juvenile Investigation Bureau

Z The largest number of recorded first-time juvenile offenses committed between 1980
and 1997 was in 1991 with a total of 298.

Z About 24 percent of the total number of first time offenses was classified as incidences
(which include run-away and missing minors).

Z Out of every 10 juvenile offenders, 8 did not use a weapon in committing their first
offense.

Z About 36.3 percent of all juvenile offenders committed two or more offenses.

Z Approximately 7 out of 10 juvenile offenders were born in the US Virgin Islands.

Z Juvenile offenders born in Puerto Rico comprised the highest percentage-41.3-of
repeat offenders among all ethnic groups.

Z Juvenile offenders born in the US Virgin Islands had the lowest percentage of any
group classified as not using a weapon during the commission of an offense.

Z Most repeat offenses 21.2 percent- occurred between 12 to 24 months after the first
offense.

Z At the time of their first offense, one out of every five-or 20 percent of- juvenile
offenders was 15 years old; most males- 18.9 percent- committed their first offense
at this age, while most females- 22.5 percent- recorded their first offense at age 14.

Z Approximately 50 percent of juvenile offenders born in the US Virgin Islands were
between the ages of 10 and 14.

Z About 48 percent of first-time offenders between the ages of 10 and 14 had mothers
who were born in the US Virgin Islands, and 43.3 percent had fathers who were born
in the US Virgin Islands.

Z Ninety-six percent of juvenile offenders were black.









Z Approximately 50 percent of juvenile offenders lived in a household with their
mother, but without their natural father.

Z Eighty-eight percent of juvenile offenders were enrolled in school at the time of their
offense.

Z About 91 percent of the offenders were attending junior high or high school at the
time of their offense, while 9 percent were in elementary school.

Z Among all offending juveniles, 40 percent were children of mothers born in the
Eastern Caribbean.

Z Seventy-three percent of mothers of juvenile offenders were employed at the time of
their arrest.


Predicting Recidivism with Juvenile Investigation Bureau Data

Z For every increase in age by one year (between 7 and 17 years) the odds of being a
repeat offender more than triples, but the chances of recidivism increase at a
decreasing rate.

Z Males are almost twice as likely as females to be repeat offenders.

Z Offenders born in the US Virgin Islands or in the Eastern Caribbean are almost twice
as likely to be recidivists as children born in the United States, whereas juveniles born
in Latin countries are about one-and-a-half times as likely to be recidivists as United
States born children.

Z Juveniles with one offense who live with a mother and a step-father or with foster
parents are nearly twice as likely to be repeat offenders as first-time offenders who
live with their natural parents (singly or together), a father and step-mother (singly or
together) or with relatives.

Z First-time male juvenile offenders, between 8 and 15 years old, born in the US Virgin
Islands and living with a mother and step-father or with foster parents, have a
relatively high risk of being repeat offenders.

Z Females most at risk of recidivism are those who are 12-years old, born in either the
US Virgin Islands, in a Latin country or the Eastern Caribbean, and who live with
their mother and step-father or with foster parents.









Department of Human Services

e Approximately 85 percent of all juvenile offenders were male.

e About 86 percent of offenders were citizens of the United States.

e Nearly 19 percent of juvenile offenders were 15 years old.

e Three out of four offenders were born in the US Virgin Islands.

e The majority of juvenile offenders -96.4 percent-were black.

e While a mother was present in the household of juvenile offenders 87.1 percent of
the time, a father was present in the household only 30.6 percent of the time.

e The majority of the mothers of offenders 61.2 percent- fell within the 40-to-49-
year age group, while the fathers were most represented in the 50 and over age
group.

e Approximately half of the mothers of juvenile offenders were born in the Eastern
Caribbean and about 60 percent of the fathers were born in the Eastern Caribbean.

e Forty-seven percent of the mothers of offenders were married and 51 percent of
the fathers were married, although not necessarily to each other.

e About 74 percent of the mothers and fathers of juvenile offenders were employed.

e Nearly one out of every two juvenile offenders were living with their mother.

e Thirty-seven percent of the natural parents of offenders were never married to
each other.

e About 23 percent of juvenile offenders lived in a household with either one or two
siblings.

e Eighty-four percent of offenders were enrolled in school; however there was no
certainty as to actual attendance and participation.

e The majority of juvenile offenders 86.7 percent- were employed.

e About sixty-two percent of juvenile offenders lived in a household with three or
more children.









SNinety-six percent of juvenile offenders did not have any children of their own.

M Approximately 58 percent of juvenile offenders had been suspended from school.

e Twenty-six percent of offenders had not used drugs or alcohol.

e About 43 percent of juvenile offenders had trouble controlling their anger and 41
percent experienced frequent arguments and fights.

K Fifty-eight percent of juvenile offenders were sexually active.


Predicting Recidivism with Department of Human Services Data

Z STT/DHS data confirm the STT/JIB findings on recidivism:

Z That each year after age seven the likelihood of a repeat offense more than
triples;

Z That males are over twice as likely as females to be charged with more than one
offense;

Z That juveniles born in the US Virgin Islands or the Eastern Caribbean are about
three time as likely to be recidivist as those born in the United States; and

Z That a first-time offender who lives with a mother and step-father or foster
parents is more likely to be charged with at least one other offense than one
who lives with natural parents (singly or together), a father and a step-mother
(singly or together), or with relatives.

Z Additionally, a first-time offender who admits to feelings of failure or worthlessness is
twice more likely to be a recidivist than one who does not.

Z Males between 8 and 16 years who were born in the US Virgin Islands and who have
low self-esteem and live with a mother and step-father or foster parents are predicted
to have a high chance of being a repeat offender.

Z A 12-year old female born in the US Virgin Islands or the Eastern Caribbean who
expresses feelings of failure or worthlessness and lives with a mother and step-father
or foster parents has a high risk being a recidivist.









ST CROIX

Youth Investigative Bureau

Z The largest number of recorded first-time juvenile offenses committed between 1980
and 1997 was in 1995 with a total of 753, and the second highest was 435 offenses in
1991.

Z About 41 percent of first-time offenses were Part I and II felonies.

Z The overwhelming majority of offenders -84.6 percent- committed only one offense;
only 15.4 percent were repeat offenders.

Z Two-thirds of the female offenders committed their first offense between the ages of
14 and 16.

Z Two-thirds of the male offenders committed their first offense between the ages 15
and 17.

Z Eighty percent of juvenile offenders were born in the US Virgin Islands.

Z About 77 percent of the mothers of juvenile offenders were born in the U.S. Virgin
Islands and approximately 69 percent of the fathers of juvenile offenders were born in
the U.S. Virgin Islands.

Z Approximately three out of every four or 75 percent of female offenders were black
and about 77 percent of the male offenders were black.

Z Only 3.7 percent of juvenile offenders were not enrolled in school.

Z Nearly 76 percent of female offenders resided in homes where the mother only was
the householder.

Z Approximately four out of every five male offenders lived in homes where the mother
only was the householder.

Z About 71 percent of first-time offenders were attending high school.









Predicting Recidivism with Youth Investigation Bureau Data

Z YIB data in St Croix suggest that:

Z Males are nearly twice as likely to be repeat offenders than females;

Z Offenders enrolled in school at the time of their first charge are between three
and four times more likely to be recidivist than those who are not enrolled;

Z Juveniles who commit their first offense while in Grade 12 have more than
twice the likelihood of being charged again than those who are not in Grade 12.

Z An offender who has a relative that has been previously charged with a
criminal offense is almost seven times more likely to be a recidivist compared
to one who does not have such a relative.


Department of Human Services

Z About 86 percent of juvenile offenders were male.

Z Approximately 94 percent of juvenile offenders were classified as a citizen of the
United States.

Z Juvenile offenders were most represented in the 16 year age category.

Z The majority of the offenders 84.6 percent- were born in US Virgin Islands.

Z About 98 percent of the juvenile offenders were black.

Z Only 27.8 percent of the offenders had a father as a householder, but 86.1 percent lived
in homes with a mother as a householder.

Z More than half of the mothers 51.1 percent- of juvenile offenders fell in the 30-to-39-
year age group, while the majority of fathers were in the 40-to-49-year age group.

Z The majority of juvenile offenders had mothers -41.1 percent-and fathers- 57.3
percent-who were married, although not necessarily to each other.

Z More than half of the juvenile offenders 51.8 percent-lived with their mother only at
the time of their first offense.









Z Forty percent of the natural parents of juvenile offenders were never married to each
other.

Z The majority of offenders live in a household with less than three siblings.

Z About 78 percent of the offenders were enrolled in school at the time of their first
offense.

Z The highest percentage of juvenile offenders- 26.8 percent- was in Grade 7 at the
time of their first offense.

Z Only 15.7 percent of the juvenile offenders were employed.

Z Nearly half of the offenders lived in a household with three or more children.

Z Approximately three percent of offenders had children.

Z About 2 out of every 10 offenders were responsible for childcare.

Z About 36 percent of the juvenile offenders expressed feelings of failure or personal
worthlessness.

Z Nearly 50 percent of offenders admitted that they had used drugs or alcohol.

Z The majority of offenders 66.5 percent- reported they had been suspended from
school at some point, while only 4.9 percent had been expelled.

Z About 29 percent of the juveniles offenders indicated that they had been hurt or
physically abused.

Z More than half- 57.9 percent- of the juvenile offenders reported themselves to be
sexually active.

Z About 9 percent of the offenders had been forced to have sex.

Z Approximately six percent of the offenders admitted that they did not know their
biological father.


xvii









Predicting Recidivism with Department of Human Services Data

St Croix/DHS data suggest that:

Z Males are over twice as likely as females to be recidivists.

Z A juvenile who lives with his or her married parents has less chances of being a repeat
offender than those who do not.

Z First-time offenders who harbor feelings of being a failure or being worthless have
almost twice the risk of being charged again compared to those who do not harbor
such negative feelings.

Z Juveniles who use drugs or alcohol have odds of recidivism that are nearly three times
higher than those who do not use these substances.

Z Males between 8 and 16 years old who do not live in a home with married parents,
take drugs or use alcohol and have low self-esteem are predicted to have a high risk of
recidivism.


XVIII









I. IAJTODmICIOA


1.1 Background to the Study

1.1.1 Historical and Social Change

Before the US Virgin Islands achieved its cur-
rent status as a United States territory, it had
gone through a series of political and adminis-
trative changes that impacted the social compo-
sition of the islands. The Spanish, Dutch, Eng-
lish and French each played a role in early
colonization before Denmark, the longest occu-
pying colonial power, took possession from
around 1665 until 1917 when the United States
purchased the territory for strategic purposes.

Political and economic changes since the 1950s
have moved the islands from being solely of
strategic military importance and a primary ag-
ricultural society, to a more diversified econ-
omy. Tourism has become the main economic
activity on St Thomas/St John, while oil and
refining and spirits manufactures are prime
economic activities on St Croix.

This more diversified economy encouraged in-
migration from United States citizens wishing
to experience the novelty of island life or from
entrepreneurs seeking economic opportunities.
Persons of nonresident status, willing to take
advantage of employment opportunities, were
brought in as bonded workers. Most emigrated
from the Caribbean, especially the neighboring
English-speaking Eastern Caribbean islands. It
was difficult for families of these bonded work-
ers to join them at any stage. These workers
had no residential rights beyond the temporary
status that a job assured. Later, the Immigration
and Nationality Act of 1965 made provisions
for the reunification of families. This Act
opened the floodgates for thousands of immi-
grants to take up residence in the Virgin Is-
lands.


Residents of African descent, who had been
brought to the territory as slaves for plantation
labor, comprised the largest sector of the popu-
lation and remained from inception mainly at
the base of the socio-economic pyramid.

Entrepreneurs of Asian descent have been
drawn to the commerce associated with tour-
ism. These immigration trends, together with a
steady and often heavy flow of immigrants
from the rest of the Caribbean, have produced a
population that is highly diverse, both ethni-
cally and culturally.

1.1.2 Some Effects of Immigration

The movement of large numbers of people with
differing ethnic and socio-economic back-
grounds into an already established community
has always been marked with accompanying
changes, resistance and social stresses as each
group contributes its own uniqueness.

The majority of US mainland migrants into the
territory have been white, and of middle-
income levels or above. These migrants tend to
reside in particular suburbs that are more afflu-
ent. This injection affected the social balance
and tension in the society, though no recent
psychological studies have been carried out to
determine the impact of these effects.

Migrants from the rest of the Caribbean, espe-
cially the Eastern Caribbean, have been usually
economically diverse, and settle in a variety of
areas. Some migrants join family and may
move into equivalent or better jobs and social
positions as those from which they have left
The majority of this group, however, seeks in-
creased employment opportunities and hope for
an improvement in the standard of living that
access to the United States is expected to pro-







INTRODUCTION


vide.

Table 1.1 shows population changes in the US
Virgin Islands between 1980 and 1997 based
on country of origin. The percent change col-
umn was calculated from 1980 and 1997 col-
umns, since more representative data were
available for these years. Over the 17-year pe-
riod, the overall population increased by 16.9
percent. The number of persons born in the US
Virgin Islands increased by 34.7 percent over
the same period. The increase from the Other
Caribbean region accounted for 42.7 percent.

In recent years there has been low, or in some
cases diminishing, immigration from some is-
lands such as Puerto Rico, St Lucia and Trini-
dad and Tobago. This is due possibly to the sig-
nificant improvement in the economies of these
islands. The British Virgin Islands, for exam-
ple, has experienced a reversal of the migration
trend in recent years with a movement of the
children of former immigrants returning there
in search of a better life. However, there con-
tinues to be increased immigration from Carib-
bean countries such as the Dominican Republic
(591.4 percent) and Dominica (67.1 percent).
The number of persons who became natural-
ized citizens increased by 62.7 percent during
this period.

A number of social and economic characteris-
tics are associated with immigrant populations,
especially those migrating in large numbers
under difficult conditions. The following is an
attempt to outline some of these characteristics
that may be applied to immigration patterns in
the US Virgin Islands.

1. The most important factor in making a deci-
sion on where to locate usually depends on
affordability of housing. New immigrants
tend to settle near urban centers in low-
quality and often substandard housing, with
easy access to affordable public transporta-
tion. They may be attracted to areas where


persons from their homeland already reside.
Availability of public housing, often with
less than desirable conditions, may also be
an option.

2. Different morals and values from those ex-
perienced in the immigrant's homeland usu-
ally prevail in the new environment and
must be learned, often rapidly, in order to
facilitate social adjustment in the adopted
community. In the absence of a systematic
or monitored introduction to the new ways
of life, children are often confused and
overwhelmed. They may be exposed to ex-
periences and behavior formerly considered
undesirable. In addition, more readily avail-
able money, illegal drugs and weapons, as
well as unsupervised opportunities for ex-
perimentation, leave children exposed and
vulnerable. At the same time, children born
of newly-arrived immigrants may be seen
as passports to helping older migrants es-
tablish themselves as belonging. This is es-
pecially seen in the case of immigrants who
speak a foreign first language and possess
weak communication skills in English. Un-
fortunately, such children may not always
be embraced quickly as individuals with
full rights into the receiving community.

3. Employment opportunities for new immi-
grants who are usually unskilled tend to be
in jobs that local residents regard as inade-
quate in wages or lacking in social status.
Employment may also be in rapidly grow-
ing sectors of the economy with an insuffi-
ciently available domestic labor force.
Wages are usually -minimum wage or be-
low in the case of undocumented immi-
grants. It may be necessary to take several
of these low-paying jobs in order to boost
wages. This situation exerts more pressure
on families, who find they must leave their
children on their own for long periods of
time while working one or two jobs. This
minimizes the time for socialization. Mi-








INTRODUCTION


Tablel.1 Place of Birth by Year: 1980, 1990, 1995, 1997


1980 1990 1995 1997 Percent


Percent Percent
Number of Total Number of Total


96,5

43,2


Place of Birth and
Citizenship

Total Persons

Born in US VI

St Croix

St John

St Thomas

Born in Puerto Rico

Born in US
Born Abroad of US
Parents

Naturalized Citizens-US

Not a Citizen of the US

Other Caribbean

Anguilla

Antigua/Barbuda

British VI

Dominica

Dominican Republic

Montserrat

St Kitts-Nevis

St Lucia

Trinidad & Tobago

Elsewhere/Not Reported


569 100.0 101,809

234 44.8 49,839

... 24,991

... 698

... 24,150

)93 5.2 3,974

)64 12.4 14,246


282

7,819

22,912

25,327

1,210

4,939

3,315

2,653

580

705

6,523

2,713


0.3

8.1

23.7

26.2

1.3

5.1

3.4

2.7

0.6

0.7

6.8

2.8


1,032

14,361

17,464

25,609

899

4,398

2,665

3,219

1,157

623

5,828

2,533


100.0

49.0

24.5

0.7

23.7

3.9

14.0

0.0

14.1

17.2

25.2

0.9

4.3

2.6

3.2

1.1

0.6

5.7

2.5


2,689 2.8 1,837

5,365 5.6 7,871


Number

109,677

55,389

26,375

675

28,329

4,468

12,689

389

18,894

17,848

34,330

1,073

5,192

4,007

4,344

2,553

621

7,932

2,603


1.8 2,699 2.5 1,667


7.7 2,801


2.6 1,861


Source: Bureau of the Census, Detailed Population Characteristics: Virgin Islands of the United States (PC80 1-D55).
Bureau of the Census, Detailed Cross-tabulations for the US Virgin Islands (CPH -L -156).
Eastern Caribbean Center, University of the Virgin Islands, 1995 Population and Housing Survey.
Eastern Caribbean Center, University of the Virgin Islands, 1997 Consumer Expenditure Survey: Population and Housing.

nors are also left up to their own devices teams that might have been available in the
regarding such behavior as the choice of home country become disrupted or nonexis-
television programs to watch or the group tent and services such as childcare may be
of friends to lime with. Family support sys- unaffordable. In the new setting, guardians


Change


1.5 -38.0

1.6 -65.3


Percent
of Total

100.0

0.5

24.0

0.6

25.8

4.1

11.6

0.4

17.2

16.3

31.3

1.0

4.7

3.7

4.0

2.3

0.6

7.2

2.4


Number

112,863

58,256

26,948

820

29,707

4,800

11,814

418

12,723

15,194

36,131

1,412

6,382

2,990

4,433

4,010

811

8,113

2,604


Percent
of Total

100.0

51.6

23.9

0.7

26.3

4.3

10.5

0.7

0.1

0.1

32.0

1.3

5.7

2.6

3.9

3.6

0.7

7.2

2.3


1980-97

16.9

34.7

7.8

17.5

23.0

-3.9

-1.3

48.2

62.7

-33.7

42.7

16.7

29.2

-9.8

67.1

591.4

15.0

24.4

-4.0


4,9

11,9







INTRODUCTION


or parents with feelings of guilt may also
allow lax, permissive behavior not usually
accepted in their homeland in order to com-
pensate the neglected child.

4. Immigrant families are often split up with
the breadwinner emigrating first, and then
gradually arranging for the rest of the fam-
ily members to join as socio-economic con-
ditions improve or as immigration policy
allows. This puts a strain on family ties and
relationships and may lead to continued
tensions. Separations often occur even after
the family is physically reunited.

5. Systems and institutions adjust slowly to
change caused by the influx of immigrants.
Thus, they may lag behind in providing so-
cial services to these families until needs
are manifested in a series of crises such as
incidents of violence. This is seen in the
case of juvenile delinquent behavior. A
heavy burden is placed on institutions pro-
viding services such as education, health-
care, law enforcement and social welfare.
The inadequate planning, reluctance to ac-
commodate the increase in numbers until it
is unavoidable, guarantees an escalation in
the rate of social problems experienced in
the community.

6. Even after institutions implement policies
and procedures to accommodate immi-
grants, there is usually a period of adjust-
ment on both sides before social acceptance
and assimilation into the new community
are achieved. The life of this period can
vary and may last for decades.

1.1.3 Risk Factors Affecting Juveniles

Researchers have identified a number of impor-
tant insights into juvenile delinquent behavior.
Youth involved in such behavior are usually
from economically stressed families and com-
munities, have histories of physical and sexual


abuse, lack educational and vocational skills
and are prone to become involved in alcohol
and other forms of drug abuse. They may be
failing or dropping out of school, and involved
in unprotected sex.

Youth agencies have, for decades, used assess-
ment instruments in evaluating high-risk youth.
The following are seven background and per-
sonal characteristics used by social workers and
counselors to predict high-risk behavior. The
higher the level or occurrence of these crucial
characteristics, the higher the child's risk-level
for delinquent behavior:

1. Age variable:
The child enters the correctional system at a
very young age.

2. Psychological variable:
Rebelliousness and identification with non-
conformity, together with a low level of
self-esteem.

3. School performance variable:
Low achievement levels, undesirable social
behavior, and truancy.

4. Home Adjustment variable:
Poor interaction between the child and par-
ents and between the child and siblings,
curfew violation, failure to respond to disci-
pline and supervision, and running away.

5. Drug and alcohol use variable:
The child starts to abuse drugs at an early
age, there is abuse of serious drugs, parents
have a history of alcohol and drug abuse.

6. Neighborhood variable:
The child lives in a neighborhood charac-
terized by disorganization, poverty, and
multiple social problems.

7. Social Adjustment variable:
The child's friends are involved in problem







INTRODUCTION


Table 1.2 Possible Indicators of Juvenile Delinquent Behavior


Status of Status of
Indicator USVI Nation

Low birth-weight babies 8% 7%

Infant deaths per 1,000 live births 11 7

Families with children <18 years old headed by a single parent 44% 27%

Death of children 1 to 14 years old per 100,000 42 26

Children in poverty 41% 20%

High school dropout rate of children 16 to 19 years old 22% 10%

Death rate of children by accident, homicide or suicide 15 to 17 82 62
years old per 100,000

Teen birth rate females 15 to 17 years old per 1,000 39 34

High school dropouts not working 16 to 19 years old 5% 10%

Violent crime rate: arrests of youth 10 to 17 years old per 310 471
100,000

Median income of families with children 0 to 18 years old $30,530 $39,700

Child abuse and neglect 0 to 18 years old per 1,000 8 NA


Source: Community Foundation of the Virgin Islands. (2000). US
Count Data Book 2000. St Thomas, VI: Author.


behavior such as delinquency, drug abuse,
truancy and disruption in school. There are
displays of sexual acting-out and associa
tion with gangs.

1.1.4 At-Risk Indicators in the US Virgin
Islands

Table 1.2 provides a comparison between chil-
dren under the age of 18 in the US Virgin Is-
lands to those children in the United States on
various possible indicators of juvenile delin-
quent behavior. Examination of the different
indicators suggests that for the most part chil-
dren in the US Virgin Islands are exposed to a


Virgin Islands Kids


higher level of risk than children in the United
States. While Table 1.2 refers to all children
under the age of 18, the findings presented be-
low are limited to juvenile offenders 7 to 17
years of age.

1.1.5 Youth Crime Trends

US Youth Crime Trends: Between 1990 and
1994 the violent crime index for juveniles un-
der 18 years old increased by 26 percent, repre-
senting four times that for individuals age 18
and over (Bartollas, 1997). Over the same pe-
riod, juveniles' arrests for murder increased by
14.6 percent compared to 4.6 percent for adults.







INTRODUCTION


A 1992 study found the arrest rate for African-
American youth was five times higher than the
rate for white youths (Bartollas, 1997, p.33).
The use of guns by juveniles in homicides in-
creased from 64 percent to 78 percent between
1987 and 1991. Juveniles' arrest for weapons
law violations during this period also increased
by 62 percent.

Status of Crime in US Virgin Islands: Accord-
ing to the 1990 Census of Population and
Housing, the St Thomas/St John community
had approximately 10,432 persons ages seven
to 17, the age group that constitutes the juvenile
category as designated by the police depart-
ment.

Crimes involving excessive violence, especially
when committed in public places such as
schools or in the full view of adults, usually
cause alarm and public outcry within the com-
munity.

During the 1990s the US Virgin Islands seemed
to be following a nationwide trend in an escala-
tion of violence by juveniles carried out on fel-
low juveniles, caretakers, teachers and others.
These assaults took place in neighborhoods by
gangs or on school premises. Violence so
seemingly deliberate, brutal at times and even
fatal, left communities searching for an under-
standing of the cause of violence and a method
of prevention.

For almost every year from 1989 to 1997, sta-
tistics from the Department of Public Safety
show a consistent number of offenses commit-
ted by juveniles in the St Thomas/St John dis-
trict. These cover the range from trespassing
and auto theft to rape and homicide.

The community has been forced to ask itself a
number of obvious, as well as not so obvious,
questions. How many juveniles are involved in
criminal behavior? Are there characteristics or
profiles that may help to identify these juve-


niles? Are the conditions to which these juve-
niles are exposed in the home, school or com-
munity contributing to delinquent behavior?
Where these conditions exist, are they in the
form of commission-the presence of some
negatively influencing factors in the young per-
sons' lives or omission, the absence of positive
factors?

Handling Juvenile Delinquency: The definition
of delinquency includes the portrayal of delin-
quent events, the reform and punishment of de-
linquents, and policy decisions about delin-
quency that occur in a particular setting. This
social setting in turn is shaped by five contexts:
historical, legal, socio-cultural, economic and
political. All are interwoven to affect the
whole.

Historically, a community's past methods of
handling juvenile delinquents influence its cur-
rent perception and methods. This past ap-
proach could well extend to an influence of his-
toric factors that may have affected juvenile
justice, rise and decline of beliefs and practices
that govern cycles of delinquent behavior.

Legal context is set by the role and jurisdiction
assigned juvenile courts, as well as constitu-
tional obligations to which courts will hold the
community responsible for dealing with youth
in trouble.

Socio-cultural context will be determined by
the effectiveness of societal institutions in in-
teraction with the delinquent. This reflects the
extent that the family, school and religious or-
ganizations, different lifestyles, peer groups,
and the process of urbanization all play in con-
tributing to juvenile problem behavior.

Economic status and the ability to succeed at
economic goals affect youth and influence their
behavior. This is especially the case in a society
where success and winning are held at such a
premium.







INTRODUCTION


Politically, although local conditions may exert
influence, national and institutional policy posi-
tions also affect youth attitudes toward crime
and shape levels of severity or leniency for
punishment (Bartollas, 1997).

1.1.6 US Virgin Islands Youth Rehabilitation
Centert

A thorough audit of the Youth Rehabilitation
Center (YRC) was carried out by the Office of
Insular Affairs (OIA). A study was also com-
pleted by the Federal Department of Justice,
Office of Juvenile Justice Delinquency Preven-
tion. The studies identified concerns such as
inadequate staffing, limited treatment and lim-
ited education opportunities for youth housed at
the YRC in St Croix.

The original facility was designed to cater to 27
male adolescents. Between 1991 and 1994,
however, the average number increased from
20, to 134. There were no policies and proce-
dures to regulate the facility. One social worker
comprised the treatment staff and there were no
health services in place. In addition, no special-
ized training was provided for those working
with adolescent offenders.

The Territory also experienced an increase, not
only in the number of juvenile delinquents, but
also in the levels of crime. This follows na-
tional and international trends.

The focus of the Department of Human Ser-
vices (DHS) in regard to the YRC since 1994
has been to put policies and procedures into
effect for the facility. DHS has worked to im-
prove community-based sentencing options,
residential programs and effective interventions
for juveniles within the community. These in-
clude early intervention programs such as Head


'Thefollowing is a paraphrased outline ofa presentation by
Mrs. Sidonie Halbert, Commissioner, Department of Human
Services (DHS), Division of Children, Youth and Families, to
the Virgin Islands Legislature on May 17, 2000.


Start, child care and after-school programs,
youth promotion programs for teenagers and
parenting that positively impacts children.
All youth at YRC are ordered to the facility by
the court. Although the Virgin Islands Code
permits Persons in Need of Supervision (PINS)
to be mixed with adjudicated youth, this is con-
trary to federal regulation. The practice puts
the availability of federal funds for the facility
in jeopardy. Adjudicated youth should not be
housed within sight or sound of an adult of-
fender. That includes adolescents transferred to
be tried as adults or sentenced as adults. In FY
1999 the facility housed 17 such adjudicated
youth with no provisions made within the facil-
ity for the required separation. In March 2000
out of 55 youths, 14 to 25 percent were adoles-
cents transferred as adults, (seven for murder
and four for rape).

All youth at the YRC facility are adolescents
ages 13 to 18. However, a judge may-under
special circumstances-retain a youth at YRC
until that person turned 19 years of age. This is
mostly for serious offenses. There were four
youth aged 18 by May, 2000. Although the fa-
cility was designed to accommodate 27 adjudi-
cated delinquents at a time, the level of occu-
pancy averaged 40 to 50. This overcrowding
was due to the presence of youth transferred as
adults who continued to be housed in the facil-
ity.

The facility was never designed to house fe-
males and never did until 1992 when two fe-
males were placed there. Subsequently, from
1994 to 1999 there were 20 or more females
remanded to the facility yearly with as many as
34 placed there in 1997. As of the report date,
56 youth were housed in the facility. Eleven of
them-20 percent-were females. In FY 1999,
the YRC served three population types: pre-
trial juveniles who were in contempt of court
orders, adjudicated juveniles and youth trans-
ferred as adults (Table 1.3).








INTRODUCTION


Table 1.3 Three Population Types
Served by the Youth Reha-
bilitation Center: FY 1999


SNumber Percent

Total 155 100

Pre-Trial Juveniles in
contempt Court Order 57 37

STT/STJ 21 37

STX 36 63

Adjudicated Juveniles 81 52

STT/STJ 38 47

STX 43 53
Youth transferred (as
adults ) 17 11

STT/STJ 8 47

STX 9 53
Note: Youth transferred as adults (both pretrial and
sentenced)
Source: Department of Human Services (2000).

In FY 1999, the youth population remanded at
the YRC had the characteristics shown in Table
1.4. In addition to those 155 juveniles re-
manded at the YRC, there were 411 who,
though not remanded at YRC, were being ser-
viced by the system (Table 1.5).

Data from the 1995 survey indicated that there
were 12,528 teenagers between ages 13 to 18 in
the Territory. The 566 in the DHS system (155
remanded at YRC, 411 not incarcerated) ac-
counted for 4.5 percent of that youth popula-
tion.

YRC Expansion: Plans are in progress for the
improvement of the facility that would include
the following:
a. Expansion of the YRC by a construction
project of three phases that will result in a


51 bed dormitory facility.
b. Development of a Pre-release Center com-
prising an eight-bed pre-release home.
c. A total of four trailers for use as office and
classroom space.
d. Security fence to replace the fence damaged
by Hurricane Marilyn in 1995.
e. Staffing included 54 Full-Time (FT) DHS
staff at YRC with an additional 3 FT and
two Part Time (PT) teachers from the De-
partment of Education.

Services and Activities: Educational activities
were provided by one FT teacher from DHS
who works all year round and five teachers
from the Department of Education (three FT;
two PT). Activities included: educational as-
sessment, GED preparation, tutoring, special
education and vocational education in the areas
of air-condition/refrigeration and upholstery.

Recreation: YRC had a full-time recreational
specialist who coordinated outdoor activities
such as basketball and softball and various in-
door activities such as board games and pool.


Treatment Services: Two full-time licensed so-
cial workers coordinated with the Intervention

Table 1.4 Characteristics of Youths
Remanded at YRC FY 1999


Characteristics Total Percent

Youths Remanded at YRC 155 100
First time admission 87 56
Single parent household 58 37
Tested positively for drug use
upon admission 57 37
School dropouts at admission 34 30
Diagnosed learning disability 20 13
Living at home at time of arrest 81 52
Arrested for Violent Crime 46 27







INTRODUCTION


Unit for intake and discharge and provided all
court work and counseling. An average 80
youth counseling sessions were provided each
month, along with six court reports and 16
court hearings. Residential counselors also
worked with juvenile residents, listening to
their concerns and helping to develop alterna-
tive ways to deal with anger, depression and
conflicts. There is an agreement between DHS
and the Virgin Islands Behavioral Services
(VIBS) to provide health services and supple-
mental mental health services to youth at YRC.

Following are some of the areas for which ser-
vices were contracted in FY 1999:
Mental Health: psychiatric evaluations, consul-
tations, psychotherapy sessions, emergency,
psychological evaluations.
General Health: physical, medical histories,
emergency visits, hospital referrals, nursing
care, medication provision, laboratory ser-
vices/tests, pharmaceutical supplies, screening.
Special Concerns: There are a large number of
youth who are testing positive for drugs, usu-
ally marijuana, upon intake. Plans have been
discussed to use monies available from the Law
Enforcement Planning Commission (LEPC) to
renovate a facility to house drug treatment pro-
grams. There was a recognized need for full
evaluations for all youth entering the facility
that may have special needs.

1.1.7 Department of Human Services

History: Aid to the poor and needy of the US
Virgin Islands was first provided from trust
funds established under the Danish colonial
administration and administered by local
"Charitable aid Councils" on each island. In
1922, under the United States Navy administra-
tion, a Department of Public Welfare was es-
tablished by order of the Naval Governor, and
upon direction of an admiral's aide. With the
transition to a civilian administration, all wel-
fare functions were administered in the Gover-


nor's office, with the first civilian Superinten-
dent of Public Welfare appointed in 1931, and
the first native appointed to that position in
1933. This person also coordinated services
provided to the Virgin Islands under the Fed-
eral Emergency Relief Act of 1933.

In 1943 the Virgin Islands Welfare Act estab-
lished the Department of Social Welfare for
providing public social welfare services on a
territorial basis. It was a legal base for the ex-
tension of federal social security programs to
the US Virgin Islands. Federal child health and
welfare services were extended to the territory
in 1947; public assistance in 1950; old age and
survivors insurance benefits in 1951; Medicaid
benefits in 1966; the Food Stamp program in
1974; and the Child Support Enforcement pro-
gram in 1975. A local medical assistance pro-
gram, which was operated by the Department
of Social Welfare since 1961, was absorbed by
Title XIX-Medicaid-and transferred to the
Department of Health in 1966. Vocational reha-
bilitation services were transferred from the
Department of Education to Social Welfare in
1969, and the Child Support Enforcement pro-
gram was transferred from Social Welfare in
the Department of Law in 1983.

The Government Reorganization and Consoli-
dation Act of 1987, VI Act 5265, established

Table 1.5 Juveniles Remanded Outside
YRC FY 1999

Juveniles Remanded
Outside YRC Total Percent
Total 411 100

Adjudicated Juveniles 337 82
STT/STJ 215 64
STX 122 36

PINS 74 18
STT/STJ 60 81
STX 14 23







INTRODUCTION


the Department of Human Services (DHS) to
provide all services for youth, children, handi-
capped, elderly and adults and families in the
low-income bracket. Programs were consoli-
dated from the former Department of Social
Welfare, Office of Community Services, Virgin
Islands Community Action Agency, Youth Ser-
vices Administration, Commission on Youth,
Commission on Aging, Commission on the
Handicapped and the Developmental Disabili-
ties Council, all of which were abolished by the
Act. In addition, three programs for the elderly
were transferred from other departments.

Mission: The mission of the Department of Hu-
man Services is to provide a sensitive, efficient,
responsive human services delivery system to
meet the needs of low-income persons, the eld-
erly, children and families, and the disabled in
the Territory. Through six program divisions,
the department provides corrective, supportive
and preventive programs, as well as advocacy
to identify human needs. DHS develops inno-
vative ways to break the cycle of dependency,
and help its clients become self-sufficient, con-
tributing members of the Virgin Islands' com-
munities.

Vision: The DHS envisions its emergence as a
department that empowers individuals and
families toward self-sufficiency through a
seamless delivery system of services.

1.1.8 Juvenile Units

Background: The Juvenile Unit of the
St Thomas/St John Police Department was es-
tablished in February, 1969. The Juvenile Unit
of the St Croix Police Department was estab-
lished on December 1, 1969. A senior detective
with extensive experience in juvenile crime
was assigned as juvenile officer to head the
unit.

The unit offers many varied services to the
community and its citizens. In accordance with


the Code of Laws of the Territory of the US
Virgin Islands, a juvenile is a person under 18
years of age.

Mission: It shall be the mission of the Juvenile
Unit to:
a) Be on the lookout for potential delinquents
and for conditions that cause delinquent be-
havior.
b) Work with other agencies to correct environ-
mental conditions that may foster delin-
quency.
c) Use preventive patrol where it has potential
for helping to control delinquency.
d) Investigate delinquency problems that lead
to juvenile offenses. This includes apprehen-
sion and prosecution of adults involved in
these cases.
e) Investigate and prosecute all complaints of
child abuse from the home when conditions
warrant.
f) Detect, and take into custody, juvenile of-
fenders for both criminal and status offenses.
g) Follow up on juvenile offenders, using refer-
ral and other dispositions as appropriate.

Duties: It is the duty and responsibility of the
Juvenile Unit to:
a) Interview all juvenile offenders and victims
in accordance with the law.
b) Investigate all criminal and status offenses
involving juveniles.
c) Determine which juvenile offenders will be
placed on diversion programs and which ju-
veniles will be referred to Family Court.
d) Prepare court complaints and provide assis-
tance to the Attorney General at hearings
and cases involving juveniles before Family
Court.
e) Investigate and prepare for prosecution,
when applicable, all cases of child abuse or
neglect. The unit may also render assistance
in child abuse or neglect cases on the island
when requested by other agencies and ap-
proved by the Head of the Juvenile Unit.
f) Prepare Uniform Crime Reports (UCR) and







INTRODUCTION


ensure proper data concerning missing or
runaway juveniles.
g) Conduct in-service training on juvenile laws
and procedures for members of the depart-
ment.
h) Assist the Bureau of Criminal Investigation
in criminal cases when the juvenile caseload
permits.
i) Conduct juvenile workshops; give lectures
and talks to adults and students on all phases
of juvenile law, crimes and abuses of the
child. These programs will include visual
aids and a realistic approach to the growing
juvenile problems in our community and na-
tion.
j) Plan, organize, sponsor, or co-sponsor sum-
mer youth information, recreational and
safety programs, i.e. Police Movie Club, Op-
eration Shipmate and bicycle rodeos. Such
police information programs will teach the
youth safety and responsibility in the home,
family and community.
k) Seek out and identify needy children and
families who deserve food and toys fur-
nished by military or civilian groups for dis-
tribution during the holiday season. This is
an on-going yearly program.
1) Work closely with school officials and
teachers to deter and suppress juvenile delin-
quency, crimes, and disorders on school
campuses.

Duties and Responsibilities of the Commander:
In addition to the general authority duties and
responsibilities of command and supervisory
personnel enumerated in Section 3.3.4 of this
chapter, it is the responsibility of the Com-
mander of the Juvenile Bureau to:
a) Develop a program of delinquency preven-
tion intended to eliminate factors that induce
criminal tendencies and conditions that pro-
mote criminal activities among juveniles.
The commander shall promote such a pro-
gram by enlisting the aid of the public, inter-
ested agencies and other divisions and bu-
reaus of the Department. The commander


shall coordinate the functions of his bureau
with that of the Territorial Court, Youth Ser-
vices Administration, Department of Wel-
fare and other organizations concerned with
the social welfare of juveniles.
b) Properly investigate and process all cases of
delinquency coming to the bureau's atten-
tion. In addition, the commander should
have members of his bureau patrol all places
where delinquency is likely to occur, to de-
tect any violations of law which involve ju-
veniles and which may contribute to vice or
delinquency. Also, the commander should
ensure that bureau members assist members
of other divisions, bureaus, sections and
units, as well as community agencies and
organizations involved in the prevention of
juvenile delinquency and youth crime.
c) Accept all juveniles taken into custody by
any member of the Department; establish an
adequate system of recording complaints,
contacts, arrests, investigations and deposi-
tions involving juveniles; and maintain a
complete and accurate file of all cases re-
ferred to the Juvenile Investigation Bureau
for action, indicating its status.
d) Collect data, discover and chart trends,
stimulate juvenile protective measures, and
work to improve general community condi-
tions. The commander should work to elimi-
nate those factors contributing to delin-
quency. Also, the commander may interpret
functions of the Department as a whole, and
the Juvenile Bureau in particular, in the sup-
pression of delinquency, when ever the com-
mander or any member of the bureau has
occasion to speak before any civil, educa-
tional or religious groups.
e) Have a thorough knowledge of the laws
regulating the handling of juveniles, and in-
struct and ensure the obedience of all such
laws by members of the bureau.
f) Report to the Deputy Chief of Police all im-
portant activities of the bureau, and keep the
deputy advised of the youth crime situation
in the district, the effectiveness of the bu-







INTRODUCTION


reau, morale, disciplinary problems or needs
of the bureau.

Types of Juvenile Offenses

There are two types of offenses peculiar to ju-
veniles:
a) Criminal Offenses: The criminal offense is a
crime against persons and/or property and is
the same offense as the adult criminal of-
fenses. Upon request, the juvenile, after re-
ceiving a waiver hearing before a Family
Court Judge, may be referred to the General
Sessions Court.
b) Status offense: the status offense is a non-
criminal offense such as a runaway or tru-
ancy.

1.2 Purpose of the LEPC Study

In the Memorandum of Understanding (MOU)
between the Virgin Islands Law Enforcement
Planning Commission (LEPC) and the Univer-
sity of the Virgin Islands (UVI) the following
were outlined as the purpose of the study:
a) Develop a statistical profile of delinquent
youth, with the primary aim of adopting a
proactive approach to the prevention and
reduction of the escalating level of crime
among youth;
b) Collect data on delinquent youth between
1987 and 1997 from the files in the Virgin
Islands Department of Human Services
(DHS) and in the Juvenile Investigation Bu-
reau of the Virgin Islands Police Department
(JIB);
c) Develop associated attributes of delinquents;
d) Determine what are the significant character-
istics which explain juvenile delinquency;
e) Measure the degree of association between
abused juveniles and delinquency;
f) Establish characteristics that distinguish a
first-time offender from a repeat offender;
g) Measure the attributes of first-time and re-
peat offenders


h) Determine the time frame between the first
and second arrest for repeat offenders;
i) Measure the effectiveness of intervention
services; and
j) Develop a prognostic evaluation model that
will assist caseworkers in determining the
likelihood of evolving from a first-time to a
repeat offender.

LEPC was charged with developing profile in-
formation on young offenders in the Territory.

Through a coordinated and cooperative effort
between the three working parties involved-
UVI, Department of Human Services and Juve-
nile Investigation Bureau-expertise and per-
sonnel from UVI were commissioned to con-
duct the study and gather the needed profile
information.

Services Performed by UVI: UVI was commis-
sioned to perform the following services in ful-
fillment of the purpose of the MOU.
a) The design of two instruments to collect spe-
cific data from the files of JIB and a sample
selected case files from DHS, primary data
items;
b) The preparation of a computer screen that
will be used in keying the data from the in-
struments;
c) Editing recorded data in the instruments be-
fore data entry takes place;
d) Generation of descriptive statistics of delin-
quents and associated characteristics from
the data;
e) Identification of appropriate statistical ana-
lytic techniques for the range of analyses to
which the data will be subjected;
f) Generation of appropriate tables for presen-
tation;
g) Application of statistical models and inter-
pretation of output; and
h) Preparation and printing of a final report for
presentation.







INTRODUCTION


1.3 Scope of the Study

Among other things, the proposed study will
seek to determine the impact of four fundamen-
tal socioeconomic issues on juvenile delin-
quency:
a) the composition of the household in which
children live-such as the number of par-
ents, grandparents and other relatives, and
the number of siblings;
b) the educational attainment of the parents;
c) the labor force behavior of the parents, espe-
cially changes in employment of mothers;
and
d) poverty status of the household.

Initially, an estimate of 4,500 case files at JIB
and 5,000 at DHS spanning the 11-year period
of 1987 to 1997 were believed to exist. At first
it was anticipated that a scientific sample would
be necessary. On closer examination, however,
it was discovered that the estimates of available
case files were highly exaggerated. The deci-
sion was then taken to utilize all the cases
available at JIB and match them with the corre-
sponding cases from DHS.

Two survey-type instruments were developed
to record data from each selected arrest or de-
linquent case. These instruments sought to
gather data based on the following categories:

From the Juvenile Investigation Bureau:
Sex
Age
Date of birth
Place of birth
Race
Birthplace of parent/guardian or caretaker
School enrollment
Education (highest level)
Employment status of guardian
Date of crime/incident
Type of crime/incident
Type of weapon used
Status (outcome/disposition).


From Department of Human Services:
General Information:
Sex
Age
Date of birth
Place of birth
Race
Date of crime/incident
Type of crime/incident
Type of weapon used
Status (outcome/disposition)
Immigration status
Mother in the household
Father in the household
Number of siblings in the household
Education (highest level)
Occupation
Marital status
Number of children
Responsible for caring for children
Type of live-in program, if any
Head of household
Marital relationship of natural parents
Education of primary caretaker(s)
Income (gross per month).

Psychological Adjustment:
Self esteem
Frequent arguments and fights with oth-
ers
Age at first contact with police
Use of drugs/alcohol
Use of drugs/alcohol before or during
school
Sexual activity
Number of different sex partners in past
year
Physically abused

Case files of juveniles in both JIB and DHS are
confidential and may be seen only by sworn
staff members. Volunteer in-house staff mem-
bers were compensated to do the data recording
after their regular work hours. The data instru-
ment for each case was edited to ensure that
items of information were assigned numeric







INTRODUCTION


codes. These codes were entered into an Inte-
grated Microcomputer Programs Systems
(IMPS) application for tabulation.

1.4 Survey Goals

With a growing number of youth becoming in-
volved in delinquent behavior in the US Virgin
Islands, it became imperative that some under-
standing of factors contributing to this problem
be investigated.

Although extensive studies have been con-
ducted on the US mainland that point to cause
and effect of juvenile delinquent behavior, to
date no study with an in-depth scientific ap-
proach has tackled this problem in the US Vir-
gin Islands. This study serves as a first in gath-
ering available data on juveniles that have
passed through the correctional/counseling sys-
tems. The data has been processed in a form
that will facilitate analysis. By examining the
data for correlations between variables, some
conclusions can be suggested regarding the fac-
tors that contribute to juvenile delinquent be-
havior.

It is anticipated that a number of indicators will


be established and a framework developed on
which to build an ongoing analysis of the prob-
lems of juvenile delinquents. It is hoped that
suggestions may be offered to assist the institu-
tions working with youth in general and juve-
nile delinquents in particular, to minimize the
problem of delinquent behavior.

It is expected that data and information pro-
duced from this study will be used by legisla-
tors with the resolve to identify resources to
underwrite necessary preventive programs. The
data can also be used by executive officers to
develop policy geared to the reduction and pre-
vention of delinquent behavior. It is also antici-
pated that information would be released to
make the public more aware of the factors play-
ing a role in delinquent behavior of the terri-
tory's youth. It is expected that this first effort
would also assist the JIB and DHS in both dis-
tricts to fine tune their instruments of data col-
lection. In so doing they can develop more effi-
cient and comprehensive instruments for the
purpose of the data collection. It is hoped that
this study will form the basis for present analy-
sis of the data gathered on juvenile delinquents,
as well as provide a benchmark for subsequent
comparison and study.










II. P1RSPrCtm IS AJ.D .METHODOL-Oq


2.1 Delinquency and Antisocial
Behavior

2.1.1 Causes and Recommendations

In the section that follows, consideration is
given to the extent to which various forms of
delinquent behavior can be predicted from ear-
lier conduct problems, the relationships among
these behaviors, and implications of this knowl-
edge for intervention focused on delinquency
reduction.

2.1.2 Development, Persistence and Path-
ways of Early Problem Behaviors2

It is relatively simple to fall prey to the notion
that juvenile crime is the manifestation of a sin-
gle latent construct such as low self-control. If
this is so, what then are the causes of, and the
constraints on, the operation of low self-
control? Ineffective child rearing-that is, fail-
ure to set clear expectations for behavior, fail-
ure to monitor children, and excessively severe
and inconsistent disciplinary practices-clearly
contributes to delinquent behavior. One theory
suggests that differences in opportunities for
personal gain may also be a cause of crime.
Three aspects of the developmental continuity
of delinquency will be discussed: different
manifestations of the problems from childhood
to adulthood, the persistence of problem behav-
ior, and the impact of behavioral catalysts on
the continuity or discontinuity of behavioral
problems.

The basic premise in the development sequence
in problem behavior is that youngsters of dif-
ferent ages have different capabilities and be-
havioral repertoires for the expression of prob-


lematic behavior. A rough approximation of a
development ordering of problematic behaviors
from early childhood to late adolescence, ac-
cording to the earliest age at which particular
behaviors manifest themselves, might conform
to the following: a difficult temperament, hy-
peractivity, overt conduct problems and aggres-
siveness, withdrawal, poor peer relationships,
academic problems, covert or concealing of
conduct problems, association with deviant
peers, delinquency leading to arrest and recidi-
vism.

A difficult temperament is one of the earliest
manifestations of a problem behavior. The
syndrome of hyperactivity-with its associated
attention deficit problems-is usually not dis-
cernible before age two or three. As soon as
the child begins to walk, this opens up many
new avenues of mischief. When children learn
to speak, verbal aggression may be added to
their behavioral repertoire. Social withdrawal
may also become apparent during preschool
years. The availability of peers in kindergarten
may generate new problems or reinforce the
aggressive problems that may already have be-
come apparent. During elementary school,
concealing problem behaviors such as truancy,
theft and association with deviant peers may
emerge. In junior high or high school, young-
sters may begin to associate with deviant peers
or begin to commit minor crimes. For the ma-
jority of children this will be a passing phe-
nomenon, but for others it will constitute a tran-
sition to the frequent commission of more seri-
ous delinquent activity. Contacts with police
tend to accelerate during early adolescence, and
so does recidivist offending.

Studies of the continuity of child behavior indi-
cate that certain behaviors function as catalysts
in that other behaviors are prone to persist


2Loeber, 1996.







PERSPECTIVES AND METHODOLOGY


when the catalyst is present; conversely, when
the catalyst is absent, the problematic behavior
tends to become less likely over time. Two ex-
amples are hyperactivity and substance use.

There is considerable evidence that hyperactiv-
ity-and associated impulsivity and attention
problems-and disruptive behaviors often oc-
cur simultaneously. Related questions that may
be posed are: To what extent is hyperactivity a
catalyst for development of disruptive behav-
ior? Is hyperactivity directly associated with
later serious outcomes, or is the relationship
conditional on the emergence of early and less
serious forms of disruptive behavior? Longitu-
dinal studies elsewhere of 6- to 16-year-old hy-
peractive boys until age 16 have shown that
those who continued to be hyperactive were far
more likely to show antisocial or substance use
disorder than those whose attention deficit dis-
order abated. It has also been reported that hy-
peractivity and conduct problems at age 8 to 10
each predicted chronic offending.

It is clear that most of the evidence indicates
the catalytic role of hyperactivity in the devel-
opment of serious disruptive behavior in child-
hood and adolescence. It should be understood
that hyperactivity is often accompanied by im-
pulsivity and attention problems, and that
probably the impulsive element in hyperactivity
is of most relevance for the associated onset of
disruptive behaviors during childhood and ado-
lescence. The following developmental se-
quence is plausible. Hyperactivity may stimu-
late an early onset of disobedience, conduct
problems, and delinquency. The risk of serious
conduct problems and delinquency may be-
come more acute when physical aggression ac-
companies disobedience at an early age.
Whether a decrease in hyperactivity is associ-
ated with a decrease in problem behavior is less
certain.

Studies have shown that the more serious the
substance use, the higher the likelihood that


individuals will engage in serious forms of de-
linquency. One longitudinal analysis of boys
from age 13 to 16 reported that the onset of
drug use-marijuana or hard drugs-and deal-
ing in adolescent males, was associated with an
increase in person-related offenses and carrying
a concealed weapon. If drug involvement and
delinquency are intertwined, does this also
mean that a decrease in drug use is followed by
a decrease in delinquent activities? There is
evidence for this: when individuals began us-
ing hard drugs less frequently, their criminal
involvement also decreased.

In brief, hyperactivity and drug involvement
may act as catalysts in the development of dis-
ruptive behaviors, but the evidence is not alto-
gether complete. Additional studies need to
demonstrate that catalysts operate independ-
ently from third factors such as impulsivity,
aggression, or peer influences.

It is common practice in criminology to classify
youngsters on the basis of their first offense,
but it can be argued that classification on the
basis of a single behavior is rather limited.
There is obviously more than one pathway to
crime. The concept of a pathway allows indi-
viduals with varying degrees of deviance to be
placed on one or more developmental trajecto-
ries. Three pathways have been identified:
first, the overt ,pal iiy', which represents an
escalation from minor aggression-such as an-
noying others, bullying-to physical fighting,
and eventually to violence as robbery, rape,
etc.); second, the covert itlmini y, consisting
first of minor covert acts-like shoplifting and
frequent lying-then property damage and then
more serious forms of theft, such as breaking
and entering; and third, the authority conflict
tini ey, which had as its first step stubborn
behavior, followed by serious disobedience and
defiance, and finally by authority avoidance
before the age of 12 in the form of staying out
late at night or truancy. Typically, the onset of
authority conflict preceded the onset of overt or







PERSPECTIVES AND METHODOLOGY


covert behavior.

In brief, the evidence points to multiple path-
ways rather than a single pathway in the devel-
opment of disruptive problem behaviors. The
knowledge of pathways is important because it
can help identify, at an early age, juveniles who
appear to be at highest risk for later maladjust-
ment and can distinguish them from youth with
minor and/or more transient problem behavior.
Pathways can also aid in the specification of
the type of problem behaviors that certain risk
groups of youth are likely to experience next.
Once pathways have been formulated, risk and
causal variables can explain why some youth
do not initiate problem behavior at all, why oth-
ers get into less serious problem behavior tem-
porarily and do not progress, and a proportion
of youth escalate in problem behavior and be-
come seriously affected.

The following section deals with some basic
theoretical ideas that have been advanced to
explain the causes of juvenile crime. Offend-
ing is a type of behavior, similar in many re-
spects to other types of antisocial behavior.
Hence, the theories, methods and knowledge of
other types of antisocial behavior can be ap-
plied to the study of crime.

2.1.3 Explanation of the Causes of Antisocial
Behavior

Human behavior is, to a considerable degree,
consistent and predictable. It has been argued
that, in general, the antisocial child tends to be-
come the antisocial teenager and the antisocial
adult, just as the antisocial adult then tends to
produce another antisocial child. This section
begins by examining some of the most impor-
tant risk factors that have been identified with
offending and antisocial behavior.3 These in-


3Risk factors are prior factors that increase the risk of
occurrence of events such as the onset, frequency, persistence or
duration of offending.


clude family influences such as poor parental
management and techniques, childhood antiso-
cial behavior, offending by parents and sib-
lings, low intelligence and educational attain-
ment, separation from parents, prenatal and
prenatal factors such as low birth weight, indi-
vidual difference factors such as high impulsiv-
ity, socioeconomic deprivation, peer influences
such as having delinquent friends, school fac-
tors, community characteristics such as social
disorganization, and situational factors
(Farrington, 1996).

2.1.4 Prenatal and Perinatal Factors

In the Caribbean, early childbearing, or teenage
pregnancy, predicts many undesirable out-
comes for the children, including low school
attainment, antisocial school behavior, sub-
stance use, and early sexual intercourse. The
children of teenage mothers are also more
likely to become offenders. Studies in the US
and England have found that teenage mothers
were associated with low-income families, wel-
fare support, absent biological fathers, used
poor child-rearing methods, and that their chil-
dren were characterized by low school attain-
ment and delinquency. Additional research has
shown that smoking during pregnancy was as-
sociated with low birth weight, small height,
and poor performance in school. Excessive
alcohol consumption during pregnancy pre-
dicted hyperactivity, low intelligence, and
speech disorders in children.

2.1.5 Hyperactivity

Hyperactivity is an important construct that
predicts later delinquency. It is associated with
restlessness, impulsivity, and a short attention
span that has been termed the hyperactivity-
impulsivity-attention deficit or HIA syndrome.
Related concepts include a poor ability to defer
gratification and a short future time perspec-
tive. It was pointed out above that hyperactiv-
ity and conduct problems between 8 and 10







PERSPECTIVES AND METHODOLOGY


years reliably predicted chronic offending, with
the proviso that hyperactivity by itself is not
necessarily associated with disruptive behavior.

2.1.6 Intelligence and Attainment

Empirical work has shown that intelligence at
age 4 significantly predicted the number of ar-
rests up to age 27. Results of a nonverbal intel-
ligence test in Cambridge showed that twice as
many boys scoring 90 or less at ages 8 to 10
were convicted as juveniles than of the remain-
der. Low nonverbal intelligence was highly
correlated with low verbal intelligence and with
low school attainment, and all of these meas-
ures predicted juvenile convictions to much the
same extent. In addition to their poor school
performance, delinquents tended to be frequent
truants, to leave school at the earliest possible
age and to take no school examinations.

Low nonverbal intelligence was especially
characteristic of the juvenile recidivists and
those first convicted at the earliest ages. It has
also been argued that high intelligence is a pro-
tective factor against offending for children
from high-risk backgrounds. Delinquents often
do better on nonverbal performance tests than
on verbal tests. The association between
school failure and offending has been demon-
strated consistently in longitudinal surveys.
They showed how the probability of offending
increased according to the presence of truancy,
school failure, and low socioeconomic status.

A more plausible explanatory factor underlying
the link between intelligence and offending is
the ability to manipulate abstract concepts.
Youngsters who are poor at this tend to do
badly on intelligence tests, and they also tend to
commit offenses, mainly because of their poor
ability to foresee the consequences of their
offending and to appreciate the feelings of vic-
tims (i.e., their low empathy). One researcher
argues that working-class, poorer parents tend
to live for the present and to have little thought


for the future, and tend to talk in terms of the
concrete rather than the abstract. A lack of
concern for the future is also linked to the con-
cept of impulsivity.

2.1.7 Parental Supervision, Discipline, and
Attitude

An exhaustive review of family factors as pre-
dictors of juvenile conduct problems and delin-
quency found that poor parental supervision or
monitoring, erratic or harsh parental discipline,
parental disharmony, parental rejection of the
child, and low parental involvement in the
child's activities (as well as antisocial parents
and large family size) were all important pre-
dictors of offending. A study conducted in
Boston reported that poor parental supervision
was the best predictor of both violent and prop-
erty crimes. Parental aggressiveness-which
included harsh discipline, shading into child
abuse at the extreme-and parental conflict
were significant precursors of violent crimes,
while the mother's attitude-passive or reject-
ing-was a significant precursor of property
crimes. Other studies also found that poor su-
pervision and discipline were consistently re-
lated to later offending, and that hostile and re-
jecting mothers when children were age 5 pre-
dicted their children's frequent use of drugs at
age 18.

It was also found that poor parental supervision
was the most important correlate of delin-
quency for girls, and that it was the second
most important for boys-after delinquent
friends. The Boston study also reported that
harsh or erratic parental discipline, cruel, pas-
sive, or neglecting parental attitude, poor super-
vision, and parental conflict predicted later ju-
venile convictions. Generally, the presence of
any of these adverse family background fea-
tures doubled the risk of later juvenile convic-
tion. Poor supervision, erratic or inconsistent
discipline, and conflict between parents are all
conducive to offending by children. It is im-







PERSPECTIVES AND METHODOLOGY


portant for parents to explain to children why
they are being punished, so that they can iden-
tify precisely the behavior that is disapproved.
One research study indicated a significant inter-
generational transmission of aggressive and
violent behavior from parents to children.
Children who were physically abused up to age
11 were significantly likely to become violent
offenders in the next 15 years.

2.1.8 Broken Homes

The theory that broken homes cause delin-
quency was inspired by psychoanalytic ideas in
1951. The research found that delinquents
were significantly more likely than comparison
children to have suffered a complete and pro-
longed separation from their mothers during the
first five years of life. It was argued that
mother love in infancy and childhood was just
as important for mental health as were vitamins
and proteins for physical health. It was thought
that it was essential that a child should experi-
ence a warm, loving, and continuous relation-
ship with a mother figure. If a child suffered a
prolonged period of maternal deprivation dur-
ing the first five years of life, this would have
irreversible negative effects, including delin-
quency.

It is important to distinguish separation from a
biological or operative parent after a loving re-
lationship has been built up, attributable to dif-
ferent causes such as death or parental dishar-
mony, from the complete lack of contact with a
biological parent. It was found that the preva-
lence of offending was high for boys reared in
broken homes without affectionate mothers and
for those reared in united homes characterized
by parental conflict, irrespective of whether
they had affectionate mothers. The prevalence
of offending was low for those reared in united
homes without conflict and-importantly-
equally low for boys from broken homes with
affectionate mothers. These findings suggest
that it is not so much the broken home that is


criminogenic as the parental conflict that
causes it. They also indicate that a loving
mother might in some sense be able to compen-
sate for the loss of a father.

2.1.9 Large Family Size

Many studies in developed countries show that
large families predict delinquency. There are
many possible reasons a large number of sib-
lings might increase the risk of delinquency.
Generally, as the number of children in a fam-
ily increases, not only does the amount of pa-
rental attention that can be given to each child
decrease, but the household will tend to be-
come more overcrowded, possibly leading to
increases in frustration, irritation and conflict.
However, one is guided by our knowledge that
in the US Virgin Islands there have been many
large (extended) families with loving parents in
which delinquency never reared its head.

A few additional comments concerning inhibi-
tions against offending are offered. First, one
researcher emphasized differences in parental
child rearing behavior as the major source of
differences in criminal tendencies. Middle-
class parents are more likely to explain to chil-
dren why they were being punished and to be
concerned with long-term character building
and the inculcation of general moral principles.
In contrast, working-class parents supervised
their children less closely and were more incon-
sistent in their use of discipline. Middle-class
parents in general tend to use love-oriented dis-
cipline, relying on withdrawal of love as the
main sanction, whereas working-class parents
used much more physical punishment. The re-
searcher contended that working-class children
committed more crimes because working-class
parents used less effective methods of sociali-
zation.

Second, two modern criminologists suggested
that people differed in their underlying criminal
tendency, and that whether a person chose to







PERSPECTIVES AND METHODOLOGY


commit a crime in any situation depended on
whether the expected benefits of offending out-
weighed the expected costs. The benefits of
offending, including material gain, peer ap-
proval, and sexual gratification, tended to be
contemporaneous with the crime. In contrast,
many of the costs of offending, such as the risk
of being caught and punished and the possible
loss of reputation or employment, were uncer-
tain and long-delayed. Other costs, such as
pangs of conscience (or guilt), disapproval by
onlookers, and retaliation by the victim, were
more immediate. The key is the extent to
which people's behavior is influenced by im-
mediate as opposed to delayed consequences.

2.2 Methodology

2.2.1 Steps in Developing the Framework of
the Study

A series of discussion meetings were convened
with representatives from the Eastern Carib-
bean Center of UVI, the Juvenile Investigation
Bureau (JIB) and the Department of Human
Services (DHS). Through these meetings, a
clearer understanding was gained of the meth-
ods, procedures and guidelines each depart-
ment/division utilized in processing juveniles
entering the system.

It was learned that youngsters 7 to 17 years
were considered juveniles and processed as
such. In rare cases, children as young as five
years appeared on file; however, these proved
to be Persons in Need of Service (PINS), vic-
tims of neglect or crime, or children lost or
abandoned; these were eventually excluded
from the study sample.

Forms used by employees in each division to
record information concerning juveniles admit-
ted into the system for delinquent behavior
formed the basis in compiling the data-
gathering instrument. These forms included the
Contact Card used by JIB as well as several


forms from the DHS that provided information
on general, personal, psychological, educa-
tional, and legal aspects of the juvenile history.

The normal process allowed an officer with the
JIB to arrest a juvenile age seven to 17 who
was involved in offensive behavior and to place
on file a Contact Card containing a detailed
description of the offender and the crime perpe-
trated. The juvenile is then referred to DHS
with a copy of the information gathered by JIB.
DHS conducts a thorough assessment of the
juvenile and makes a decision about the needs
of each juvenile, as well as the affected com-
munity.

Two survey instruments were compiled by the
Eastern Caribbean Center in collaboration with
JIB and DHS. Since all juveniles must first pass
through JIB, that became the starting point of
data-gathering. To ensure confidentiality and
avoid disclosure to unauthorized persons,
names of juvenile delinquents were adhered to
the data-gathering instrument by a removable
label to facilitate identification by JIB and DHS
data gatherers. Each case was then identified by
a serial number consisting of four ordinate dig-
its and a first offense date for identification.

When JIB employees completed the recording
of data for all cases they were passed on to
DHS where the process was continued with in-
formation from files of juvenile delinquents.
Name-identifying labels were removed from
the data-gathering instrument before delivery to
the Eastern Caribbean Center for processing.

It was discovered that the total number of cases
existing on file at JIB did not match total cases
at DHS. This was due at least in part to the ex-
istence of some cases where juveniles with first
offenses that were considered minor in nature
were placed on file at JIB, counseled and re-
leased into the care of family members but not
passed on to DHS for further processing.







PERSPECTIVES AND METHODOLOGY


Data gathering commenced at JIB in the
St Thomas-St John district and data were re-
corded for all cases on file for the years 1987 to
1997 inclusive. When all cases were recorded
from both JIB and DHS, the data-gathering in-
struments were edited manually, and numeric
codes were inserted in preparation for data en-
try. An Integrated Micro-computer Processing
Systems (IMPS) software program was used to
produce data dictionaries and data application
screens.

2.2.2 The Data Collection Instrument

Two data-collection instruments were devel-
oped: the Contact Card and Arrest Record and
the Department of Human Services Form to
meet the special requirements of data available
at JIB and DHS, respectively. Each instrument
sought to capture information on each juvenile
related to the first offense.

The Contact Card and Arrest Record consisted
of 10 items that required basic identification,
background, family information as well as ar-
rest data and offenses. Arrest data and offenses
included total number of offenses recorded
throughout the juvenile's contact with JIB, as
well as arrest date and offense, type of crime
and weapon used for up to six incidents:
namely, the first three and the last three of-
fenses.

Forms for both districts carried the same items
with the exception of an additional item on past
involvement of family members in criminal
activities for the St Croix Contact Card and
Arrest Record

The DHS data collection instrument consisted
of 29 items that sought information on the juve-
nile's ethnicity, educational progress, political
status, social background, economic status, and
psychological adjustment. Information was also
gathered on the status and outcome of the six
offenses detailed in the JIB Contact Card and


Arrest Record.

In processing juvenile delinquents in the dis-
trict of St Croix, JIB and DHS did not carry out
the same type of collaboration as seen in the
district of St Thomas/St John. The St Croix dis-
trict did not track juveniles across the two divi-
sions, as did St Thomas/St John. The district of
St Croix also added an item to test the extent to
which a juvenile offender was acquainted with
their biological father.

2.2.3 Selection of the Population

In the St Thomas/St John district, the popula-
tion consisted of all juveniles that were on JIB
files dated 1987 to 1997. Some of the earlier
offenses of these cases extended to years prior
to 1987

These data records were then matched at DHS,
and filled out for each case by JIB personnel
using available data. Both data forms-without
personal identifying information-were then
handed over to the Eastern Caribbean Center.

Overall, there was no strict consistency in the
process of data-gathering between
St Thomas/St John and St Croix districts. In
St Croix, there was no match between JIB and
DHS in the data-gathering process, since there
were no systems in place to track each case
from one division to the other.

2.2.4 Data Collection Procedures

In both divisions, employees who pledged con-
fidentiality and were familiar with the forms
used to store data on juveniles' case files, were
employed temporarily as data gatherers as well
as supervisors. Preliminary briefings and train-
ing sessions were conducted with employees to
ascertain the need for accuracy and integrity in
the data gathering process. Guidelines for fill-
ing out the data collection instruments were
prepared to ensure uniformity and clarity of the







PERSPECTIVES AND METHODOLOGY


process. These included a description of simple
steps in the data-gathering process. In some
cases, suggestions were given as to where cer-
tain pieces of data could be located in the divi-
sion's forms.

The designated supervisor in each division was
responsible for overall quality of the job, to be
confirmed by thorough checks, before delivery
of the finished product to the Eastern Caribbean
Center.

In the St Thomas/St John district, data-
gathering instruments completed by JIB were
turned over to the supervisor at DHS for con-
tinuation of the data-gathering process. This
ensured that the names of juvenile offenders
remained confidential. On St Croix after com-
pletion, the data-gathering instruments were
forwarded to the Eastern Caribbean Center in-
dependently.

2.2.5 Data Availability and Deficiencies

The sources from which data were gathered
were the manual case files kept by juvenile-
arresting officers at JIB and social workers at
DHS. The forms used by these persons to re-
cord information were designed to serve the
needs of these agencies, while processing juve-
niles through various stages of the system. Data
used in this study were therefore limited to
what was available and the data gathering in-
struments could only take advantage of data
already present on the forms.

In addition, whereas the basic information on
identification of juveniles was routinely filled
in, there seemed to have been diligence in cap-
turing data on areas such as social background,
educational levels, ethnicity and psychological
adjustment. A significant number of cases had
items with missing data.


2.2.6 Data Analysis Procedures

Data entered into the IMPS Integrated Micro-
computer Processing System storage files were
examined for mistakes or incongruence by
tabulating frequency tables and making adjust-
ments where necessary. New frequencies and
cross tabulations were produced and key items
or indicators were looked at to identify pres-
ence of relationships. For example, household
type was compared with number and type of
juvenile delinquent; place of birth of juvenile
delinquent and number of offenders.

2.2.7 Organization of the Report

In general, background information in this re-
port is presented for the Territory with specific
references to districts. There is comparative
section on the Territory which also indicates
differences between districts. However, the sec-
tion on Survey Findings is presented by district,
since the data collected for
St Thomas/St John were not the same variables
as those for St Croix. Whereas
St Thomas/St John could track cases at JIB
with their counterparts at DHS, this was not
possible on StCroix.


In this report each data set
St Thomas/St John and St Croix is each
sented and analyzed separately by district.


for
pre-


2.2.8 Limitations of the Study

Ideally, a study that seeks to develop profiles or
characteristics that could be used as red light or
risk identifiers in the prevention of juvenile de-
linquency would be more comprehensive if the
attributes of delinquents, as well as those of
non-delinquents, were examined and a com-
parison made of the results. Since this study
focuses only on the delinquents, it is difficult to
make categorical statements as to possible
cause and effect without an adequate available
control group.







PERSPECTIVES AND METHODOLOGY


The population of the study is restricted to the
files that were available in JIB and DHS divi-
sions at the time of data-gathering. In the
St Thomas/St John district a number of files
were identified for which data were recorded at
JIB, but did not have corresponding matches at
DHS. Cross tabulation and analysis of data for
certain items in both sets of data were limited
to items with cases presented in both sets of
data.

At times there were inadequate or insufficient
data available on division forms used as source
information to fill out data collection instru-
ments for the study. For some items, a signifi-
cant number of cases had missing data, having
been left vacant by persons originally or per-
sons handling that particular case.

In the district of St Croix, on the other hand,
since JIB and DHS data were each gathered


without a cross reference for cases, data could
only be processed in separate JIB and DHS
batches. No direct connection could be made
between these groups of data. It was therefore
not possible to process the St Croix data to
track relationships between variables as in-
depth as that carried out for the
St Thomas/St John district.

On the higher level, since the method of data-
gathering differed in each district, it was not
possible to combine the data sets for the two
districts and look at relationships as they oc-
curred for the territory as a whole.

These differences in data gathering occurred
due to different forms and methods of data col-
lection and storage within the two districts, as
well as a difference in the relationship and sup-
port systems practiced within districts by JIB
and DHS.











III. PRsmU ar Of AAvAmrs


St Thomas Data: Description and Analysis

3.1 Juvenile Investigation Bureau Data
Variables

Data collection commenced at Juvenile Investi-
gation Bureau (JIB) in the St Thomas/St John
district.3 Data were recorded for all cases on
file for the years 1987 to 1997. The data collec-
tion instrument developed was the Contact
Card and Arrest Report (CCAR). Data were
recorded for each arrest and the following char-
acteristics were covered in the CCAR:
Sex
Age
Date of birth
Place of birth
Race/ethnic origin
Person with whom living
Birthplace of mother and father
Enrollment status
Education (highest level)
Employment status of guardian
Date of crime/incident
Type of crime/incident
Type of weapon used

3.1.1 St Thomas Data Description: Juvenile
Investigation Bureau

Year of First Offense-There were 2,348
cases of first offenses in the JIB data set (Table
3.1.1). On the average there were about 198.7
first offenses annually between 1987 and 1997.
There was a gradual increase annually from
1987 to 1991 with the exception of 1990. The
period 1992 to 1997 was mixed: there were de-
clines in 1992 and 1993, an upsurge in 1994,
declines again in 1995 and 1996, and an in-

3All future references to the St Thomas/St John district are
shortened to St Thomas.


crease in 1997. The smallest number of of-
fenses was 104 in 1996, and the highest was
298 in 1991. This suggests a wide range in the
annual number of offenses. It is not certain
whether this disparity was due to a difference
in record keeping or whether it was real.

Type of First Offense-In Table 3.1.2, Part I
Felony accounted for 21.8 percent of first of-
fenses. Part I Felony includes robbery, feloni-
ous assault, burglary, grand larceny, homicide,
rape and auto theft.
Part II Felony accounted for 11.5 percent of all
offenses. These include possession of con-

Table 3.1.1 Year of First Offense


Year of Offense Number Percent
Total 2348 100.0
Before 1980 4 0.2
1980 1 0.0
1981 8 0.3
1982 13 0.6
1983 9 0.4
1984 15 0.6
1985 9 0.4
1986 22 0.9
1987 180 7.7
1988 219 9.3
1989 259 11.0
1990 220 9.4
1991 298 12.7
1992 243 10.3
1993 178 7.6
1994 204 8.7
1995 135 5.7
1996 104 4.4
1997 146 6.2
After 1997 81 3.4








PRESENTATION ANALYSIS


Table 3.1.2 Type of First Offense


Year of Offense Number Percent

Total 2335 100.0
Part I Felony 509 21.8
Part II Felony 268 11.5
Misdemeanor 309 13.2
Incidence 548 23.5
Other 701 30.0
Part I Felony: Robbery, Felonious assault,
Burglary, Grand larceny, Homicide, Rape,
Auto theft.
Part II Felony: Possession of controlled
substance unlicensed firearm stolen
use of vehicle, Fraud, Forgery.
Misdemeanor: Aggravated assault and battery,
Petit larceny.
Incidences: Run-away minor, missing minor.


trolled substances, unlicensed firearms, stolen
property, unauthorized use of vehicle, fraud,
and forgery. Misdemeanor accounted for 13.2
percent, which includes aggravated assault and
battery, and petit larceny. Incidences accounted
for 23.5 percent, and this category includes run-
away minors and missing minors. "Others" ac-
counted for 30.0 percent.


Table 3.1.3 Weapon Used in First Offense


Weapon Number Percent

Total 2332 100.0
Gun 64 2.7
Knife 90 3.9
Hand 149 6.4
Bottle 23 1.0
Penis 11 0.5
None 1828 78.4
Other 167 7.2


Weapon Used in First Offense-The weapon
categories were gun, knife, hand, bottle, penis,
none and other. In 78.4 percent of the cases no
weapon was used (Table 3.1.3). Hand was used
in 6.4 percent of the cases, followed by knife
(3.9 percent), gun (2.7 percent), bottle (1 per-
cent), and penis (0.5 percent).

Total Number of Offenses-The majority-or
about close to two-thirds (63.7 percent)-of the
juveniles had only one offense. About one in
six, or 17.1 percent, had two offenses, and 7.3
percent had 3 offenses. The percentages for
more than three and more than six were 8.2
percent and 3.1 percent, respectively (Table
3.1.4).

Table 3.1.4 Total Number of Offenses


Offense Number Percent


Total


2340
1491
401
170
92
58
43


100.0
63.7
17.1
7.3
3.9
2.5
1.8


11 0.5


11 0.5
16 0.7
6 0.3
3 0.1
2 0.1
6 0.3
3 0.1
1 0.0
1 0.0
1 0.0







PRESENTATION ANALYSIS


3.1.2 St Thomas: Juvenile Investigation
Bureau Data Relationships

Total Number of Offenses by Place of
Birth-Table 3.1.5 represents the total number
of offenses cross-classified by place of birth.
Total numbers of offenses were grouped in two
categories: one offense, and two or more of-
fenses. In the case of juveniles born in the
United States or other countries, three-fourths
more of them had only one offense; less than
one quarter of them had two or more offenses.
For juveniles having birthplaces in the Virgin
Islands (VI), Puerto Rico (PR) and the Eastern
Caribbean (EC), the percentages for two or
more offenses were higher-37.6 percent, 41.3
percent and 36.0 percent, respectively.

Type of First Offense by Juvenile's Place of
Birth-The data in Table 3.1.6 cross-classify
'type of first offense by place of birth'. In the


Virgin Islands, Part I Felony accounted for 21.7
percent of the cases (Table 3.1.2). In those
cases where the birthplace was Puerto Rico, or
the Eastern Caribbean, it accounted for 24.4
and 24.2 percent, respectively. Both of these
were above the overall percentage for the Vir-
gin Islands.

Juveniles with a US birthplace had 15.2 percent
of their offenses in the Part II Felony category,
which was much higher than that-11.0 per-
cent-for the Territory overall. Puerto Ricans
had the lowest percentage of misdemeanor
cases-4.4 percent-while native-born Virgin
Islanders had the highest-14.3 percent. Virgin
Islanders had the lowest percentage-21.1 per-
cent-of Incidences.

Type of Weapon Used by Place of Birth-
Table 3.1.7 reveals that the overwhelming ma-
jority of the offenses-77.3 percent-were not


Table 3.1.5 Total Number of Offenses by Place of Birth


Place of Birth
Virgin Islands Puerto Rico United States Eastern Caribbean Other
Offenses Number Percent Numberl Percent Number Percent Number Percent Number Percent
Total 1667 100.0 46 100.0 165 100.0 422 100.0 34 100.0
1 Offense 1040 62.4 27 58.7 125 75.8 270 64.0 27 79.4
2 or more 627 37.6 19 41.3 40 24.2 152 36.0 7 20.6

Table 3.1.6 Type of First Offense by Place of Birth

Place of Birth
Type of Virgin Islands Puerto Rico United States Eastern Caribbean Other
Offense Number Percent Number Percent Number Percent Number Percent Number Percent
Total 1668 100.0 45 100.0 165 100.0 422 100.0 34 100.0
Part I Felony 362 21.7 11 24.4 30 18.2 102 24.2 3 8.8
Part II Felony 184 11.0 4 8.9 25 15.2 51 12.1 4 11.8
Misdemeanor 238 14.3 2 4.4 16 9.7 50 11.8 3 8.8
Incidences 352 21.1 15 33.3 53 32.1 114 27.0 13 38.2
Other 532 31.9 13 28.9 41 24.8 105 24.9 11 32.4







PRESENTATION ANALYSIS


Table 3.1.7 Type of Weapon Used by Place of Birth

Place of Birth
Virgin Islands Puerto Rico United States Eastern Caribbean Other
Weapon Number Percent Number Percent Number Percent Number Percent Number Percent

Total 1676 100.0 46 100.0 165 100.0 422 100.0 34 100.0
Gun 49 2.9 0 0.0 2 1.2 12 2.8 1 2.9
Knife 69 4.1 1 2.2 3 1.8 16 3.8 1 2.9
Hand 117 7.0 4 8.7 9 5.5 18 4.3 1 2.9
Bottle 16 1.0 0 0.0 1 0.6 6 1.4 0 0.0
Penis 10 0.6 0 0.0 0 0.0 1 0.2 0 0.0
None 1295 77.3 38 82.6 138 83.6 328 77.7 27 79.4
Other 120 7.2 3 6.5 12 7.3 41 9.7 4 11.8


committed with weapons. In cases where a
weapon was used by juveniles born in the US
Virgin Islands, the most commonly used ones
were by Hand (7.0 percent), Knife (4.1 per-
cent), and Gun (2.9 percent).

Duration Between First and Second Of-
fense-The mode-or the category with the
largest number of occurrences-for the dura-
tion between the first and second offense was
"One to less than two years." One out of every
five repeat offense took place between 12 and
24 months after the first offense (Table 3.1.8).

Age at First Offense by Sex-Table 3.1.9 in-
dicates juvenile offense levels were relatively
very low between the ages of seven and nine-
two percent or below. The number increased to
3.2 percent for 10-year-olds. Offense rates con-
tinued to increase with age, climbing to 12 per-
cent for 13-year-olds, and remaining high until
age 17. The peak of activity occurred at 15, the
age at which 19.5 percent of juveniles commit-
ted their first offense. The data suggest a posi-
tive association between age and the number of
juveniles committing their first offense.

When the data are broken down by gender, it is
observed that a pattern of increased activity oc-


curred for both sexes between the ages 13 to
17. The largest proportion of males committed
their first offense between the ages 13 and 17,
while this occurred for females between ages
13 to 16. By age 17 females had dropped to 9.9
percent of cases having first offenses, while the
males were still relatively active, with 13.1 per-
cent committing their first-time offense at this
age. These data would seem to suggest that fe-
males abandoned their delinquent activity at an

Table 3.1.8 Duration Between First
and Second Offense

Duration Number Percent
Total 824 100.0
< 1 Month 104 12.6
1 to < 2 Months 75 9.1
2 to < 3 Months 43 5.2
3 to < 6 Months 103 12.5
6 to < 12 Months 155 18.8
1 to < 2 Yrs 175 21.2
2 to < 3 Yrs 75 9.1
3 to < 4 Yrs 50 6.1
4 to < 5 Yrs 18 2.2
5 Yrs and Over 26 3.2








PRESENTATION ANALYSIS


Table 3.1.9 Age at First Offense by Sex


Total Male Female
Age Number Percent Number Percent Number Percent
Total 2282 100.0 1633 100.0 649 100.0
Age 7 22 1.0 18 1.1 4 0.6
Age 8 37 1.6 32 2.0 5 0.8
Age 9 46 2.0 40 2.4 6 0.9
Age 10 73 3.2 66 4.0 7 1.1
Age 11 108 4.7 91 5.6 17 2.6
Age 12 177 7.8 127 7.8 50 7.7
Age 13 274 12.0 193 11.8 81 12.5
Age 14 430 18.8 284 17.4 146 22.5
Age 15 446 19.5 308 18.9 138 21.3
Age 16 391 17.1 260 15.9 131 20.2
Age 17 278 12.2 214 13.1 64 9.9


earlier age than did males. This observation
coincides with other child development con-
cepts that assign higher levels of physical and
social development to females at an earlier age
than to males.

Place of Birth of Juveniles by Sex-Table
3.1.10 indicates that the majority of juvenile
offenders were born in the US Virgin Islands-
73.3 percent. There were 12.4 percent of
juvenile offenders born on an Eastern
Caribbean Island and 7.2 percent of juvenile


delinquents who were born in the US. Female
offenders were slightly less likely to by born in
the US Virgin Islands or and Eastern Caribbean
Island and slightly more likely to be born in the
US than male offenders.

Age at First Offense by Place of Birth-In
Table 3.1.11 juveniles from all places of birth
under the age of 10 appeared to have a low par-
ticipation in delinquent behavior compared to
the older age groups. Significant increases in
the rate occurred between ages 10 to 14 for all


Table 3.1.10 Place of Birth of Juvenile by Sex


Place of Birth
Total
United States
US Virgin Islands
Eastern Caribbean
Other Caribbean
Puerto Rico
Other


L|


Total
Number Percent
2288 100.0
165 7.2
1676 73.3
283 12.4
90 3.9
46 2.0
28 1 2


Male
Number Percent
1627 100.0
106 6.5
1202 73.9
206 12.7
62 3.8
31 1.9
20 1.2


Female
Number Percent
640 100.0
57 8.9
460 71.9
73 11.4
28 4.4
14 2.2
8 1 3


I


_r


I


C


-~--~- -~-- -~-- -~-- -~-- -~--







PRESENTATION ANALYSIS


Table 3.1.11 Age at First Offense by Juvenile's Place of Birth

Place of Birth
Virgin Islands Puerto Rico United States Eastern Caribbea Other
Age Number Percent Number Percent Numberl Percent Numberl Percent Numberl Percent

Total 1642 100.0 45 100.0 160 100.0 422 100.0 34 100.0
Under 10 yrs 93 5.7 1 2.2 4 2.5 16 3.8 1 2.9
10 to 14 yrs 769 46.8 19 42.2 80 50.0 183 43.4 17 50.0
15 to 16 yrs 587 35.7 20 44.4 58 36.3 167 39.6 10 29.4
Over 16 yrs 193 11.8 5 11.1 18 11.3 56 13.3 6 17.6

groups. The peak age group for which first time Age at First Offense by Father's Place of
delinquent activity occurred for all juveniles, Birth-The frequencies among the age catego-
regardless of place of birth, was between ages ries in Table 3.1.13 for place-of-birth of moth-
10 and 14, with a falling off at the "Over 16" ers are not widely different from those in 3.1.12
age category. for place-of-birth of fathers. However, while
the percentages in the age 10 to 14 group are
Age at First Offense by Mother's Place of higher for mothers that were bor in the Virgin
Birth-In Table 3.1.12 the same pattern ap- Islands and Puerto Rico, the frequency in the
pears as for Table 3.1.11 above. First-time de- same age category is higher for fathers bor in
linquent activity remained low for juveniles the US than for fathers born elsewhere.
under age 10 for all place of birth groups, but
increased significantly among the 10- to 14- Juvenile's Race by Sex-Table 3.1.14 shows
year-old age group. There was a higher rate of that of the total number of juveniles who com-
first-time offenses in the 10 to 14 age group- mitted a first offense, 96.0 percent were black,
50 percent-in the case of mothers bor in while 1.7 percent were white. Very small per-
Puerto Rico, compared with the other place of centages were classified as Asian and Other-
birth groups. This would seem to suggest that a 0.2 and 2.0 percent, respectively. These per-
greater proportion of juveniles who have a centages were generally consistent for males
mother born in Puerto Rico become involved in and females.
delinquent activity at an early age.

Table 3.1.12 Age at First Offense by Mother's Place of Birth

Mother's Place of Birth
Virgin Islands Puerto Rico United States Eastern Caribbean Other
Age Number Percent Number Percent Number Percent Number Percent Number Percent
Total 688 100.0 52 100.0 105 100.0 1167 100.0 32 100.0
Under 10 yrs 39 5.7 2 3.8 12 11.4 87 7.5 1 3.1
10 to 14 yrs 332 48.3 26 50.0 44 41.9 490 42.0 10 31.3
15 to 16 yrs 251 36.5 17 32.7 35 33.3 433 37.1 15 46.9
Over 16 yrs 66 9.6 7 13.5 14 13.3 157 13.5 6 18.8







PRESENTATION ANALYSIS


Table 3.1.13 Age at First Offense by Father's Place of Birth


Father's Place of Birth
Virgin Islands Puerto Rico United States Eastern Caribbear Other
Age Number Percent Number Percent Number Percent Number Percent Number Percent
Total 478 100.0 45 100.0 110 100.0 1141 100.0 30 100.0
Under 10 yrs 51 10.7 2 4.4 11 10.0 75 6.6 1 3.3
10 to 14 yrs 207 43.3 19 42.2 55 50.0 487 42.7 11 36.7
15 to 16 yrs 168 35.1 17 37.8 31 28.2 423 37.1 12 40.0
Over 16 yrs 52 10.9 7 15.6 13 11.8 156 13.7 6 20.0


In a 1995 survey of the general population of
St Thomas and St John, among 5- to 19-year-
olds, blacks accounted for 89.4 percent of the
total, while whites accounted for 4.3 percent,
and those of other race represented 6.3 percent.
Although the age categories do not match ex-
actly-the age group of the data set is 7 to 17-
this gives a broad general comparison of the
situation.

Juvenile's Guardian by Sex of Juvenile-The
data in Table 3.1.15 show the majority of juve-
nile delinquents-54.8 percent-lived with
their mother alone. The second highest percent-
age-21.1 percent-lived with both mother and
father. The remainder of the guardian catego-
ries ranged between 0 and 10.4 percent. Pro-
portionally, eight times as many lived with
"mother and stepfather" as opposed to "father


and stepmother"-3.3 percent to 0.4 percent,
respectively.

When comparisons are made by gender, both
male and female categories of juvenile offend-
ers followed similar patterns within similar
ranges. All of the remaining guardian catego-
ries experienced relatively low proportions of
the total number of offenders. One can assume
with some confidence that this is quite likely a
reflection of the representation of these groups
in the general Virgin Islands' population.
Namely, the assumption is that the majority of
children in this age group live with mothers
alone, and the second largest group live with
both natural parents.

School Enrollment at Time of Arrest-Table
3.1.16 indicates that only a small percentage of


Table3.1.14 Juvenile's Race by Sex


Total Male Female
Race Number Percent Number Percent Number Percent

Total 2346 100.0 1666 100.0 657 100.0
Black 2252 96.0 1602 96.2 629 95.7
White 39 1.7 26 1.6 13 2.0
Asian 4 0.2 2 0.1 2 0.3
Other 47 2.0 33 2.0 13 2.0







PRESENTATION ANALYSIS


Table 3.1.15 Juvenile's Guardian by Juvenile's Sex

Total Male Female
Guardian Number Percent Number Percent Number Percent

Total 2,346 100.0 1,666 100.0 657 100.0
Mother & Father 496 21.1 366 22.0 123 18.7
Mother 1,285 54.8 914 54.9 359 54.6
Father 137 5.8 105 6.3 31 4.7
Mother and Step Father 77 3.3 63 3.8 14 2.1
Father and Step Mother 9 0.4 5 0.3 4 0.6
Step Mother 3 0.1 2 0.1 1 0.2
Step Father 5 0.2 3 0.2 2 0.3
Foster Parents 23 1.0 8 0.5 15 2.3
Relatives 245 10.4 157 9.4 85 12.9
Other 46 2.0 32 1.9 14 2.1


Table 3.1.16 School Enrollment


Status Number Percent

Total 972 100.0
Enrolled 855 88.0
Not Enrolled 117 12.0


offenders-12 percent-were not enrolled in
school at the time of arrest. This suggests that
being enrolled in school is not a deterrent to
youngsters committing offenses. A more com-
prehensive picture might be gained if data were
available for analysis on levels of school and
class attendance, and even achievement meas-
ures.

School Type at Time of Arrest-Table 3.1.17
shows data on the school level at which juve-
nile delinquents were enrolled at the time of
their first offense. These data seem to correlate
with previous sets of data that reflected age and
first offense. The data show that a relatively
low percentage of juveniles committed first of-


Table 3.1.17 School Type

School Number Percent
Total 799 100.0
Elementary 74 9.3
Jr. High 360 45.1
High 365 45.7
fenses in elementary school-9.3 percent-
increasing to 45.1 percent in Jr. high school,
and 45.7 in high school.

Mother's Place of Birth by Juvenile's Sex-
Table 3.1.18 indicates that the majority of juve-
nile offenders-39.8 percent-were children of
mothers bor in the Eastern Caribbean. The
second highest proportion-35 percent-were
children whose mothers were natives of the US
Virgin Islands. The percentage of juvenile de-
linquents with mothers bor in the other Carib-
bean islands or island group was 14.9 percent.

Mother's Employment by Sex of Juvenile-
Table 3.1.19 shows that 72.9 percent of juve-
nile offenders had mothers who were em-







PRESENTATION ANALYSIS


Table 3.1.18 Mother's Place of Birth by Sex of Juvenile

Total Male Female
Place of Birth Number Percent Number Percent Numberl Percent
Total 2,014 100.0 1,438 100.0 559 100.0
United states 108 5.4 76 5.3 32 5.7
US Virgin Islands 705 35.0 501 34.8 198 35.4
Eastern Caribbean 801 39.8 577 40.1 218 39.0
Other Anglo Caribbean 300 14.9 214 14.9 84 15.0
Puerto Rico 53 2.6 36 2.5 14 2.5
Other 47 2.3 34 2.4 13 2.3

Table 3.1.19 Mother's Employment by Sex of Juvenile

Total Male Female
Status Number Percent Numbed Percent Number Percent

Total 1,979 100.0 1,435 100.0 527 100.0
Employed 1,443 72.9 1,062 74.0 370 70.2
Unemployed 536 27.1 373 26.0 157 29.8


played. Male offenders were slightly more
likely to have employment mothers-74.0
percent-than were females-70.2 percent.
The finding that most of the mothers of juve-
nile offenders were employed may suggest that
employed mothers were away from the home
for extended hours, especially at times when
children returned home from school. These data
do not indicate if any of the employed mothers
held more than one job or worked overtime.
These factors would potentially add to a
mother-child separation, and leave unsuper-
vised children exposed and vulnerable to in-
volvement in delinquent behavior.

3.1.3 Predicting Recidivism with the
St Thomas Juvenile Investigation
Bureau Data

3.1.3.1 Approach

One of the primary purposes of the study iden-


tified above was to adopt a proactive approach
to the study of juvenile delinquent behavior. A
persistent issue that is of significance to the De-
partment of Human Services (DHS) relates to
the task of trying to determine if a first-time
offender will commit additional violations that
will repeatedly result in his or her arrest. The
data from the Contact and Arrest Reports show
that during the study period, about 36 percent
of all arrested juveniles had been arrested more
than once. Of the total, about 12 percent of the
juveniles had been arrested 4 or more times. It
is the relatively high percent of repeat offenses
that argues for closer study of this phenome-
non.

It is evident why DHS has a primary interest in
knowing what are the factors that might explain
why some first-time offenders-male or fe-
male-never commit another offense to pre-
cipitate an arrest, while other youth persist in
patterns of delinquent behavior that cause them







PRESENTATION ANALYSIS


to be arrested several times. A keen under-
standing of this phenomenon becomes even
more critical because DHS social caseworkers
realize that juvenile repeat offenders often ma-
ture into adult criminals.

The question to be resolved might be stated
another way: what is the combination of char-
acteristics that might help to predict if a first-
time offender might become a recidivist or re-
peat offender, that is, one showing a tendency
to relapse into a previous mode of delinquent
behavior?

Statistical methods provide one technique
known as logit modeling that can help to pro-
vide answers to the question just posed. In the
sections below, this statistical method will be
applied to some of the demographic character-
istics of juveniles that were recorded in the sur-
vey data, age, place of birth, race, place of birth
of parents, school enrollment status and grade
at time of arrest. These data were collected
from both the Juvenile Investigation Bureau
(JIB) and DHS offices in St Thomas/St John
and in St Croix.

[Note: Readers who are not interested in the
statistical methodology or procedure of logisti-
cal analysis may skip sections 3.1.3.2 to
3.1.3.4].

3.1.3.2 Methodology

Although logit modeling or logistic regression
is a powerful statistical analytic technique, it
may be presented in relatively simple terms for
the non-statistical reader. An elementary de-
scription of the logit model is that it is one that
objectively examines in a numerical way the
relationship between a dependent variable and
one or a set of independent characteristics. In
this case, the method determines the chances of
a first-time offender becoming a recidivist, and
it also helps to measure the strength of the im-
pact of the demographic characteristics on the


chance of being a repeat offender. Further, it
also enables the analyst to determine the order
of the importance of the characteristics that are
significant in their influence upon reverting to
offensive behavior.

A concept that is commonly used in this kind of
analysis is that of odds. This term is commonly
used by gamblers and is easily understood by
most persons. The odds indicate the relative
number of chances of falling into one of two
categories on some characteristic of interest.
For example, if in a population of couples liv-
ing together-married and unmarried-the
chance of divorce or break-up is 0.4 and the
chance of remaining married or living together
is 0.6, then the odds of getting divorced or
splitting up would be 0.4/0.6 = 0.67. This indi-
cates that couples are two-thirds as likely to
break-up as they are to stay together.

3.1.3.3 Definition of Variables

Dependent Variable: The primary outcome
variable or characteristic of interest for each
individual in the data set is a dichotomous vari-
able (consisting of two groups). A juvenile was
classified as a one-time offender if he or she
were recorded as having been arrested one and
only one time. The juvenile was classified as a
recidivist if he or she were arrested two or more
times within the period of study. The one-time
offender is the referent group.

Independent or Explanatory Variable: These
variables or characteristics are included in the
study because they have been determined by
prior preliminary analysis to exert relatively
strong influence on the outcome variable.
These include age, sex and ethnicity of the ju-
venile, and the primary person with whom the
youth lives. Other variables that were consid-
ered, but not included in the final predictive
model were place of birth of the mother and
father, whether the mother and father were
presently occupied, school enrollment status of







PRESENTATION ANALYSIS


the youth at the time of first arrest, and grade
level at the time of first arrest. Additionally,
the type of offense and the type of weapon used
were also included in a preliminary test of vari-
ables that might exert statistically significant
impact on the outcome variable. Even though
logic suggested that certain demographic vari-
ables should be included in such a model, only
those that met statistical significance were re-
tained in the final estimating equation.

i) Age: The variable age is simply the age
of the juvenile at the time of first arrest.
Using the actual age instead of a set of
age categories affords an analysis of the
changing chances or odds of recidivism
as he juvenile moves across the age
spectrum that is measured-7 to 17
years.

ii) Age Squared: This is a theoretical vari-
able that captures the essence of the re-
lationship between age and recidivism
in that as juveniles mature, the chances
of additional offenses level off. That is,
the chances of recidivism decrease with
age at a decreasing rate. (Age squared
is therefore age multiplied by age.)

iii) Gender: This variable is dichotomous
and refers to the sex of the juvenile.
Female is the referent category.

iv) Ethnicity: Juveniles were classified into
four categories based on the juvenile's
place of birth at the time of first arrest.
The groups were: born in the Virgin
Islands, born in Puerto Rico or other
Hispanic country, born in the Eastern
Caribbean, and born in the Mainland
United States. This latter is the referent
group or the group with which the oth-
ers were compared.

v) Guardian: The primary person with
whom the juvenile was living at time of
first arrest was grouped into one of two


classifications.
step father" or
other juveniles
"Other parent".
erent group.


One was "mother and
"Foster parents". All
fell into the category
This latter was the ref-


3.1.3.4 The Model

Let it be assumed that R, is recidivism or repeat
offending, A, is the age of each juvenile, AS, is
age squared, S, is sex or gender, E, is ethnicity
and G, is primary guardian. It may further be
assumed that p, is the chance-or conditional
probability-that recidivism is high, and 1 -p,
that the chance that it is low, given A,, AS, S,,
E,, and G,. The logistic regression model for
the log odds of a recidivist is:

Log(p/(1-p))= log R, = a + 83(A,) + 32(AS,)
+ ,3(S,) + 34(E,) + B,(G,)

where log 9/(l-p)) is simply the log odds of
becoming a recidivist, given the explanatory
variables, and the betas represent the change in
the log odds due to unit increments in the val-
ues of the predictors (Menard, 1995). In sim-
pler terms, the model can be written as:


Prob(event)


1 / (1 + e-z)


where Z is the familiar linear combination

Z = Bo +B1X1+B2X2 + ... + BXp.

3.1.3.5 Estimates of the St Thomas Juvenile
Investigation Bureau Data

Numerical results from estimation of the logis-
tic regression model are shown in Table
3.1.3.1. While the total number of cases ini-
tially included in the analysis was 2,346 of-
fenders, only 2,275 were used to estimate the
results shown in the table. Seventy-one of the
cases had at least one item of information miss-
ing, and had to be excluded.
At the bottom of the main part of the table is a
row that indicates the Likelihood ratio chi-







PRESENTATION ANALYSIS


Table 3.1.3.1 Parameter Estimates in Logistic Regression Predicting Recidivism:
St Thomas, Juvenile Investigation Bureau Data

Parameter Standard Pr> Adjusted
Variable Estimate Error Chi Square Odds Ratio
(1) (2) (3) (4)
Intercept -6.5287 1.2110 <0.0001

Age 1.1726 0.1871 <0.0001 3.230

Age Squared -0.0521 0.0072 <0.0001 0.949

Sex (1=Male) 0.3220 0.0525 <0.0001 1.904

Ethnicity
USVI Borna 0.2044 0.0941 0.0298 1.879

Puerto Rico or Dominican 0.0083 0.2045 0.9675 1.545
Republic Borna

Eastern Caribbean Borna 0.2138 0.1157 0.0647 1.897

Guardian (1=Mother and 0.3201 0.1068 0.0027 1.897
Stepfather or Foster Parents)b
Likelihood Ratio Chi Square = 196.1936 df= 7 Pr > Chi Square < 0.0001
a Compared to US Born.
b Compared to living with mother and father, mother only, father only, father and step mother,
step mother only, step father only, foster parents, or relatives.
Notes: N=2275. Odds ratios greater than 1.0 indicate greater odds of recidivism.


square = 196.1936. This is a test of the hy-
pothesis that all of the explanatory variables
have coefficients of 0. Since the associated p-
value is less than 0.01, it is concluded that the
hypothesis is substantiated and that at least one
of the coefficients for the independent variables
in column (1) is not 0.

It was defined above that the odds of an event
occurring are the ratio of the chance that it will
occur to the chance that it will not occur.
When a parameter estimate-as shown in Table
3.1.3.1-is positive, the odds ratio will be
greater than one, which means that the odds of
the occurrence of the event are increased. If
the parameter estimate is negative, the odds
ratio will be less than 1, and the odds of the
event are decreased.


The variable Age has a positive sign on the pa-
rameter estimate-shown in the second column
of Table 3.1.3.1. This positive coefficient for
Age indicates that the likelihood of recidivism
increases with Age. The negative sign associ-
ated with the variable Age Squared means that
the chances of repeated offenses increase with
age at a decreasing rate. Thus, having been ar-
rested once, the odds of relapsing into offensive
behavior with a one-year increase in age are
about three times as likely as not being arrested
again. This information is derived from the
adjusted odds ratio of 3.23 for Age shown in
the last column. This means that for a one-year
increase in age, the odds of becoming a recidi-
vist are more than three times that of not be-
coming a repeat of offender. The table shows
that the negative estimate for Age Squared-
0.0521-produces an odds ratio that is slightly







PRESENTATION ANALYSIS


less than 1.0. Because the relationship between
age and recidivism is nonlinear, this variable
helps to dampen the effect of age as the juve-
niles become older.

The estimate for Sex is positive with an odds
ratio of 1.904. This indicates that the chance of
being a recidivist, with other variables held
constant, is about twice as high among males as
it is for the base case, or for females.

Four ethnic groups were included in the model,
and are explained in the lower half of Table
3.1.3.1. Three ethnic types are compared to the
base case of US born. The odds ratio for born
in the US Virgin Islands and born in the East-
ern Caribbean are 1.879 and 1.897, respec-
tively. The odds ratio for juveniles born in
Puerto Rico or the Dominican Republic is
smaller than those of the other two, with a
value of 1.545. It is noted, however, that the
column headed 'Pr > Chi Square' shows that
this variable greatly exceeds the 0.05 level of
significance, and is therefore not considered
statistically substantive. This is probably due
to the relatively small number of observations
that fell into this category when compared to
the others.

A first-time juvenile offender born in either the
US Virgin Islands or in the Eastern Caribbean,
with an odds ratio of 1.879 and 1.897, respec-
tively, is almost twice as likely to become a
recidivist as a stateside-bom first-time of-
fender. A Hispanic juvenile first-time offender
that is one born in either Puerto Rico or the Do-
minican Republic, is only slightly less likely
than either Virgin Islands-born juveniles or
Eastern Caribbean youth to become a repeat
offender.

The guardian or the primary person with whom
the offender was living at the time of first arrest
was classified into two groups. The first group
included a juvenile who was living with his or
her mother and a stepfather, or with foster par-


ents. The second group consisted of all other
guardians, primarily mother and father together
or mother alone, father alone, father and step-
mother, stepmother, stepfather, relatives or
other kinds of living arrangements.

The odds ratio for the variable Guardian of
1.897 indicates that the odds of being a recidi-
vist, with other variables held constant, are
about 90 percent more for first-time offenders
living with a mother and a stepfather or with
foster parents, than it is for a first-time offender
in some other kind of living arrangement.

3.1.3.6 The Prediction of Recidivism

The logit model can be used effectively to pre-
dict or forecast the probability or the chances
that a juvenile first-time offender will become a
repeat offender. If a juvenile has predicted
chances of offense status that are less than 0.5,
it is said that his or her chances of recidivating
are 0. If the prediction is greater than 0.5, the
chances of becoming a recidivist are 1.0, i.e., it
is highly likely that the juvenile will be a re-
cidivist. Similarly, if predicted odds are greater
than 1.0, then the probability or chances of the
outcome will be, or correspond to a probability,
greater than 0.5. And if the predicted odds are
less than 1.0, then the probability will be less
than 0.5. Therefore, in cases where the predic-
tion of the odds is less than 1.0, or the probabil-
ity is less than 0.5, it can be said that the juve-
nile will likely not have another arrest.

The data in Table 3.1.3.2 present various com-
binations of offender characteristics and the
associated chances of recidivism. The table
shows the parameter estimates of the variables
taken from Table 3.1.3.1, as well as various
juvenile profiles. The predicted chances of the
first profile shown, Al, are those of an offender
who is eight years old, is male, was born in the
US Virgin Islands, and who lives with a mother
and stepfather, or with foster parents. Column
(10) of the table shows that the predicted odds








PRESENTATION ANALYSIS


Table 3.1.3.2 Predicted Probabilities in Logistic Regression: St Thomas, Juvenile Investigation Bureau Data

Proba-
Coefficients: a AgeFO AgeSq Sex Ethnicityl Ethnicity2 Ethnicity3 Guardian Odds ability
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
Parameter
Estimates: -6.5287 1.1726 -0.0521 0.3220 0.2044 0.0083 0.2138 0.3201

Juvenile Profiles:


9 81 1 1 0 0 1
10 100 1 1 0 0 1
11 121 1 1 0 0 1
12 144 1 1 0 0 1
13 169 1 1 0 0 1
14 196 1 1 0 0 1
15 225 1 1 0 0 1
16 256 1 1 0 0 1


Al
A2
A3
A4
A5
A6
A7
A8
A9
A10
All
A12
A13
A14
A15
A16
A17
A18
A19
A20
A21
A22


1.439
1.918
2.302
2.490
2.427
2.131
1.686
1.202
0.772
0.447
0.873
1.759
1.995
1.775
1.759
1.775
1.202
0.716
0.988
0.873
0.639
1.211


0.590
0.657
0.697
0.713
0.708
0.681
0.628
0.546
0.436
0.309
0.466
0.638
0.666
0.640
0.638
0.640
0.546
0.417
0.497
0.466
0.390
0.548


Age of juvenile at first offense.


AgeSq = The square of the age of juvenile at first offense.
Sex = Sex of the juvenile (1 = male, 0 = female).
Ethnicityl = Virgin Island Born.
Ethnicity2 = Puerto Rican Born.
Ethnicity3 = Eastern Caribbean Born.
Ethnicity4 = US Mainland Born.
Guardian = Person with whom living (1 = mother and stepfather or foster parents, 0 = other guardian).


12 144 1 0 1 0 1
12 144 0 0 0 1 1
12 144 0 1 0 0 1


Variables:
AgeFO =







PRESENTATION ANALYSIS


Figure 3.1.3.1 Predicted Chances of Recidivism: St Thomas JIB
For a USVI-Born Male Living With Mother and Stepfather



0.6




0.4


0.2 --


Al A2
Note: See text for profile characteristics


A3 A4 A5 A6
Juvenile Profile


A7 A8 A9 A10


corresponding to this combination of character-
istics are greater than 1.0-1.439-and there-
fore have a prediction of recidivism of 0.590.
Because this value is greater than 0.5, this may
be interpreted that an 8-year-old male offender,
born in the Virgin Islands and living with his
mother and stepfather or with foster parents is
highly likely to become a repeat offender.

Profiles Al to A10 show outcomes for juve-
niles with identical characteristics except for
age. Thus it is noted that as age increases for
the foregoing combination of attributes, the
chances of recidivating increase at a decreasing
rate until age 11 (for profiles Al to A4). This
is illustrated graphically in Figure 3.1.3.1. It
shows that the chances of repeat offenses in-
crease at a decreasing rate until the chances
peak with A4 at age 11, and then decrease at an
increasing rate to age 17 thereafter. While the
attributes of A8 indicate that this 15-year-old
juvenile will be a recidivist, the chances are


slightly less than that of a 14-year-old. How-
ever, by the time the juvenile male with the as-
sociated characteristics reaches age 16, his pre-
dicted odds drop below 1.0 and his probability
is less than 0.5, thus indicating he will not be a
repeat offender.

Juvenile profile Al l-of a 15-year-old US Vir-
gin Islands born male living with either both
parents, or a mother alone, or a father, or a fa-
ther and stepmother, a stepmother alone, a step-
father, or with relatives-has predicted odds of
less than 1.0 and a probability of 0.466. This
juvenile is therefore predicted to be a non-
repeat offender, and is not likely to be arrested
again up to age 17. The profile of juvenile A12
is that of a 12-year-old female offender born in
the US Virgin Islands who lives with her
mother and stepfather or with foster parents.
Her chances are predicted to be about 64 per-
cent, and this indicates that she is highly likely
to commit at least one other offense and be ar-










PRESENTATION ANALYSIS


rested before age 18. Other profiles in the table
show similar combination of characteristics,
and the odds and probabilities can be inter-
preted similarly (Figure 3.1.3.2).


3.2 Department of Human Services Data
Variables

The following data items were collected from
case files at the St Thomas Department of Hu-
man Services (DHS):
Sex
Age
Date of birth
Place of birth
Race
Date of crime/incident
Type of crime/incident
Type of weapon used
Status (outcome/disposition)
Immigration status
Mother in the household
Father in the household


Number of siblings in the household
Education (highest level)
Occupation
Marital status
Number of children
Responsible for caring for children
Type of live-in program, if any
Head of household
Marital relationship of natural parents
Education of primary caretaker(s)
Income (gross per month)
Psychological Adjustment
Self esteem
Frequent arguments and fights with others
Age at first contact with police
Use of drugs/alcohol
Use of drugs/alcohol before or during
school
Sexually active
Number of different sex partners in past
year
Physical abuse


Figure 3.1.3.2 First-time Arrests of Juveniles in St Thomas
Juvenile Investigation Bureau Data



20
I Male
I Female

15


10







7 8 9 10 11 12 13 14 15 16 17
Aae







PRESENTATION ANALYSIS


There were 707 records in the St Thomas dis-
trict DHS data set. As reported earlier these
consisted of all juveniles who were on DHS
files during the period 1987 to 1997. Many of
the tables below will reflect differing totals be-
cause of missing data.

3.2.1 St Thomas Data Description: Depart-
ment of Human Services

Juvenile's Sex-According to Table 3.2.1 a
large proportion of the total juvenile offend-
ers-84.5 percent were males compared to 15.5
percent who were females. The age category
here refers to 7- to 17-year olds: as a compari-
son, the 5-to-19-year-olds in the total
St Thomas/St John population had a 50.4 to
49.6 percent male-female ratio (as estimated in
the 1995 Population and Housing Survey). This
over representation of males involved in delin-
quent behavior is a common and well-known
phenomenon, though the reasons for this occur-
rence are still not as clear.

Immigration Status-In Table 3.2.2 US citi-
zens made up the overwhelming majority of
juvenile offenders-86.2 percent-whereas
juveniles of resident alien status accounted for
just 13.8 percent. The closest corresponding
age group in the US Virgin Islands general
population for which data are available is 5 to
19; US citizens and resident aliens or perma-
nent residents in this group made up 88.4 and
10.8 percent, respectively (1995 Population and
Housing Survey).


Table 3.2.1 Juvenile's Sex


Juvenile
Sex Number Percent

Total 691 100.0

Male 584 84.5

Female 107 15.5


Table 3.2.2 Immigration Status

Juvenile
Immigration Status Number Percent
Total 623 100.0
US citizen 537 86.2
Resident Alien 86 13.8



Juvenile's Age-Delinquent activity remained
at a low rate between ages 7 to 9 (Table 3.2.3).
At age 13 there was a relative picking up of
activity and this increased dramatically be-
tween ages 14 to 17 with a peak at ages 14 and
15. Although delinquent activity at age 17 re-
mained in the high-level category, the begin-
ning of a falling off can be observed.

Juvenile's Place of Birth-The data in Table
3.2.4 show that 74 percent of juveniles commit-
ting first-time offenses were born in the US
Virgin Islands, 4.3 percent on the US mainland,
17.2 percent in Eastern Caribbean islands, and
2.6 percent in Puerto Rico.

Juvenile's Race-Table 3.2.5 shows that 96.4
percent of juveniles involved in delinquent be-


Table 3.2.3 Age of Juvenile


Juvenile
Age in Years Number Percent


Total


693 100.0
35 5.1


10-12


14.7


80 11.5


14 123 17.7
15 129 18.6
16 118 17.0
17 106 15.3







PRESENTATION ANALYSIS


Table 3.2.4 Juvenile's Place of Birth

Juvenile
Place of Birth Number Percent

Total 690 100.0
United States 30 4.3
Virgin Islands 514 74.5
Puerto Rico 18 2.6
Eastern Caribbean 119 17.2
Others 9 1.3


havior were black and 1.1 percent were white.
Data on the age group 5 to 19 in the general
population (1995 Population and Housing Sur-
vey) indicated that 89.4 percent of
St Thomas/St John's population was black, 4.3
percent white, and 6.3 percent fell in the
'Other' category. These figures suggest that a
disproportionately larger number of blacks be-
came involved in delinquent behavior while
whites and other groups were underrepresented
based on their size in the general population.

Table 3.2.5 Juvenile's Race


Juvenile
Race Number Percent
Total 700 100.0
Black 675 96.4
White 8 1.1
Asian 2 0.3
Others 15 2.1


Household Status of Parents-Of the cases in
Table 3.2.6 reporting data on the presence of
parents in the household from which juvenile
delinquents reside, 87.1 percent of mothers
lived in the household compared to 30.6 per-
cent of fathers. This supports a general concept
in discourse on the causes of juvenile delin-
quency. It suggests that in households where
either parent is absent, the tendency is to foster
a less supportive and stable environment that
permits a higher level of risk to juveniles be-
coming delinquent. Unfortunately, data are not
available on the presence of mothers and fa-
thers in the household for the general popula-
tion.

Age of Parents-In Table 3.2.7 fathers and
mothers of juvenile delinquents appear to have
generally fallen in different age categories. The
majority of mothers-61.2 percent-fell in the
40 to 49 age group. The greater proportions of
fathers (over half) were in the "50 years and
above" age group. Fathers tended to be some-
what older than mothers.

Birthplace of Parents-The data in Table
3.2.8 show that only 35.7 percent of the moth-
ers and 28.5 percent of the fathers were born in
the US Virgin Islands. The percentage for the
US mainland-born was 5.2 and 6.1 percent for
mothers and fathers, respectively. The majority
of the parents were born in the Eastern Carib-
bean. The leading islands were St Kitts and Ne-
vis with 18.6 and 17.6 percent, respectively if
the data on place of birth of parents in the case
files are compared on general population, a no-
table feature emerges. In the general popula-


Table 3.2.6 Household Status of Parents


Mother Father
Household Presence Number Percent Number Percent

Total 635 100.0 604 100.0
In the household 553 87.1 185 30.6
Not in the household 82 12.9 419 69.4








PRESENTATION ANALYSIS


Table 3.2.7 Age of Parents


Mother Father
Age in years Numberl Percent Number Percent

Total 237 100.0 105 100.0
14-19 2 0.8 0 0.0
20-29 5 2.1 0 0.0
30-39 85 35.9 12 11.4
40-49 145 61.2 40 38.1
50 & above 0 0.0 53 50.5
Table 3.2.8 Place of Birth of Parents

Mother Father
Place of Birth Number IPercent Number Percent
Total 614 100.0 555 100.0
United States 32 5.2 34 6.1
Virgin Islands 219 35.7 158 28.5
Puerto Rico 17 2.8 11 2.0
Eastern Caribbean 320 52.1 325 58.6
Others 26 4.2 27 4.9

tion (1995 Population and Housing Survey)
mothers and fathers born in St Kitts and Nevis
were only 12.5 and 11.7 percent, respectively.
In the case of Antigua and Barbuda, in the gen-
eral population, percentages were 7.7 and 9.2
for mothers and fathers, respectively; whereas
among juveniles in the data set, they were 10.6
and 11.8, respectively.

Marital Status of Parents-Table 3.2.9 shows
data on marital status for parents of those juve-
nile delinquents. Available data show that in
the highest category, 46.8 percent of mothers
and 50.7 percent of fathers were married at the
time of the juvenile's first offense, ("Now Mar-
ried"), though not necessarily to each other.
The next two categories for mothers were
"Never Married," 28.2 and "Divorced" 12.3
percent. (Here again, divorce not necessarily
between the two parents of the juvenile). For
fathers, the next two categories were the same:
"Never Married" was 24.3 percent and


"Divorced", 12.7 percent. A higher "Separated"
rate, 9.4 percent, existed for fathers of juveniles
than it did for mothers of juveniles, 8.5 percent.

Employment Status of Parents-Of the cases
in Table 3.2.10 reporting data on the employ-
ment status of parents, 74.0 percent of mothers
were employed compared to 73.5 percent of
fathers.

Juvenile's Guardian-Of the reported cases
of juvenile offenders that had information on
the person or persons with whom they lived,
48.5 percent lived with their mother, 4.9 per-
cent with their father, 23.4 percent with mother
and father, 8.4 percent with relatives and 10.5
percent with a mother and a stepfather (Table
3.2.11). These data suggest that more juvenile
delinquents who lived with their mother alone
ended up becoming delinquent more than juve-
niles in the other categories. With almost half
of the juvenile cases living in households with
a mother alone, the association between these
two variables in the data is evident.


Table 3.2.9 Marital Status of Parents

Mother Father
Marital Status Number I Percent Number Percent

Total 365 100.0 276 100.0
Now married 171 46.8 140 50.7
Widowed 15 4.1 8 2.9
Divorced 45 12.3 35 12.7
Separated 31 8.5 26 9.4
Never married 103 28.2 67 24.3

Table 3.2.10 Employment Status of Parents

Employment Mother Father
Status Number Percent Number Percent
Total 592 100.0 431 100.0
Employed 438 74.0 317 73.5
Unemployed 154 26.0 114 26.5







PRESENTATION ANALYSIS


Table 3.2.11 Juvenile's Guardian


Guardian Number Percent
Total 629 100.0
Mother and Father 147 23.4
Mother 305 48.5
Father 31 4.9
Mother and Stepfather 66 10.5
Father and Stepmother 6 1.0
Foster parents 4 0.6
Relatives 53 8.4
Other 17 2.7


Marital Relationship of Natural Parents-
Data on current marital status of natural parents
highlight categories for the natural parents of
juvenile delinquents. At time of the first of-
fense "Never Married" and "Divorced" catego-
ries represented 36.6 and 12.4 percent, respec-
tively, making up about half of the juvenile de-
linquents. The proportion for the category
"Married and Living Together" was 28.7 per-
cent. The categories "One Parent Deceased"
and "Separated" accounted for 12.1 and 10.2
percent respectively (Table 3.2.12).



Table 3.2.12 Marital Relationship of Natural
Parents

Current status Number Percent

Total 421 100.0

Married and living together 121 28.7
Separated 43 10.2

Divorced 52 12.4
Never married 154 36.6

One parent deceased 51 12.1
Both parents deceased 0 0.0


Siblings in the Household-Data in Table
3.2.13 indicate that juvenile offenders were
most represented in households with one or two
siblings present. These two categories com-
bined represented 45.2 percent of all juvenile
offenders. In households with three siblings,
21.6 percent of offenders were represented. In
contrast, only 10.1 percent of juvenile offend-
ers resided in households with four siblings and
only 4.2 percent and 4.8 percent of juveniles
lived in households with five or six or more
siblings, respectively.

School Enrollment-Table 3.2.14 indicates
that 84.4 percent of juvenile delinquents were
enrolled in school. Unfortunately, data indicat-
ing actual attendance at school and class par-
ticipation were not part of the data set. Al-
though the majority of juveniles involved in
delinquent behavior were enrolled in school,
that does not mean they attended school, par-
ticipated in and benefited from class activities.

Table 3.2.13 Siblings in the Household

Number of siblings Number Percent
Total 523 100.0
No siblings 74 14.1
One siblings 118 22.6
Two siblings 118 22.6
Three siblings 113 21.6
Four siblings 53 10.1
Five siblings 22 4.2
Six or more siblings 25 4.8
Note: Median number of siblings in the household is 2.


Table 3.2.14 School Enrollment


Enrollment Status Number Percent
Total 546 100.0
Enrolled 461 84.4
Not Enrolled 85 15.6







PRESENTATION ANALYSIS


Table 3.2.15 Current or Highest Grade Completed

Grade Number Percent
Total 689 99.6
Grade 1 1 0.1
Grade 2 212 30.6
Grade 3 0 0.0
Grade 4 10 1.4
Grade 5 7 1.0
Grade 6 20 2.9
Grade 7 105 15.2
Grade 8 111 16.0
Grade 9 107 15.5
Grade 10 59 8.5
Grade 11 34 4.9
Grade 12 23 3.3


It is important to note that 15.6 percent were
not enrolled in school, although the law stipu-
lates that children below the age of 16 must be
enrolled. Attendance at school and class is nei-
ther enforced nor easily enforceable.

Current or Highest Grade Completed-
Table 3.2.15 indicates that juveniles involved
in delinquent activities were represented across
Grades 1 through 12. Children in Grades one
through six were involved at low levels of juve-
nile delinquent behavior except Grade two
which was surprisingly high-30.6 percent.
The peak was between juveniles in Grades
seven through nine, with a range between 15.2
and 15.5 percent of the juvenile delinquents
taking part. The proportion of juveniles in-
volved in delinquent behavior began to de-
crease at Grade 10 and continued through to the
higher grades. It should be remembered that
these data represented the educational status at
time of first offense and cannot predict the edu-
cational progress after that offense.


3.2.2 St Thomas: Department of Human
Services Data Relationships

Juvenile's Employment Status-Table 3.2.16
indicates that the majority of juveniles involved
in delinquent behavior were unemployed-86.7
percent. Here again it could help to have fig-
ures showing the proportion of this age group
of juveniles who are employed and unem-
ployed in the general population in order to get
a clearer picture.

Current Job-Table 3.2.17 suggests that at the
time of the first offense, the juvenile delin-
quents were employed in a variety of job types
and no particular activity seems to have had a
monopoly on their time. It is assumed that these
are part-time jobs but there is no way of know-
ing weekly or monthly duration. Another point
to remember is that only one out of 10 reported
current jobs due to the fact that majority of
them (86.7 percent) reported unemployed.



Table 3.2.16 Juvenile's Employment Status


Currently Working Number Percent

Total 429 100.0
Employed 57 13.3
Unemployed 372 86.7


Table 3.2.17 Current Job


Job Number Percent
Total 68 100.0
Restaurant or Hotel 11 16.2
Construction 6 8.8
Store clerk 5 7.4
Mechanic shop 5 7.4
Gas station attendant 3 4.4
Other 38 55.9








PRESENTATION ANALYSIS


Table 3.2.18 Number of Children in the Household

Children Number Percent
Total 485 100.0
0 5 1.0
1 90 18.6
2 88 18.1
3 111 22.9
4 95 19.6
5 36 7.4
6-10 53 10.9
11-14 4 0.8
15 or more 3 0.6


Children in the Household-Table 3.2.18
seems to suggest that households with one, two,
three or four children were more likely to have
delinquent juveniles-between 18.6 and 22.9
percent-than households with five or more
children. It is possible that the presence of more
children provide examples of alternate behavior
and persons with whom youngsters may associ-
ate and communicate with. It also may limit
negative outside influences on their behavior
and activities.

Juvenile's Children-Table 3.2.19 shows that
a small proportion of juvenile delinquents-4.4


Table 3.2.19 Juvenile's Children

Parental Status Number Percent
Total 405 100.0
Children 18 4.4
No Children 387 95.6


Table 3.2.20 Juvenile Responsible
for Childcare

Childcare Number Percent
Total 396 100.0
Yes 14 3.5
No 382 96.5


percent-had children of their own. Juveniles
who were responsible for childcare-Table
3.2.20 show a slightly lower proportion-3.5
percent-than those with children. This respon-
sibility did not necessarily refer to children
owned by the juvenile delinquent, and could
have been a younger sibling or relative.

Presence of Risk Indicators-Each of the
variables in Table 3.2.21 holds some signifi-
cance for the juvenile's psychological well-
being. The presence of one or more of these
factors could contribute to the undermining of


Table 3.2.21 Juvenile's Psychological Assessment

Yes No
Total Number Percent Numberl Percent
Felt like a failure or worthless 171 44 25.7 127 74.3
Had trouble controlling anger 196 84 42.9 112 57.1
Had frequent arguments and fights 186 76 40.9 110 59.1
Ever been suspended from school 213 123 57.7 90 42.3
Ever been expelled from school 174 40 23.0 134 77.0
Used drugs or alcohol 167 43 25.7 124 74.3
Getting hurt (physically abused) 124 16 12.9 108 87.1







PRESENTATION ANALYSIS


the juvenile's self esteem and self worth. The
influence of any one or more of these variables
at any level usually suggests negative impacts
on behavior. There were relatively high per-
centages in areas such as had trouble control-
ling anger (42.9 percent), had frequent argu-
ments and fights( 40.9 percent), ever been sus-
pended from school (57.7 percent), used drugs
or alcohol (25.7 percent), felt like a failure or
worthless (25.7 percent), getting hurt
(physically abused)-(12.9 percent), ever been
expelled from school (23 percent).

Juvenile's Sexual Activity-Table 3.2.22 indi-
cates that 56.7 percent of the first time offend-
ers were sexually active. This variable is often
identified as an indicator associated with delin-
quent behavior.

Forced Sex-Table 3.2.23 indicates that 6.5
percent of the juveniles with first offenses were
forced to have sex. Looked at along with the
previous table on Juvenile's Sexual Activity,
the data suggest that a large proportion of juve-
niles opt to participate in sexual activity freely.

Juvenile's Sex Partners-In Table 3.2.24 the
number of juveniles (32) for whom data are
available here is relatively low. About 60 per-
cent reported one or two sex partners but 9.4

Table 3.2.22 Juvenile's Sexual Activity

Sexual Activity Numbed Percent
Total 164 100.0
Sexually Active 93 56.7
Not Sexually Active 71 43.3

Table 3.2.23 Juvenile Forced to Have Sex

SNumber Percent

Total 123 100.0
Forced 8 6.5
Not Forced 115 93.5


Table 3.2.24 Juvenile's Sex Partners

Sex Partners Number Percent
Total 32 100.0
1 14 43.8
2 5 15.6
3 5 15.6
4 2 6.3
5 3 9.4
6-10 1 3.1
11 or more 2 6.3


percent reported six or more sex partners. The
low number of cases for this variable was
probably due to the reluctance of juveniles to
be forthcoming with data that would obviously
be considered very sensitive.

Pre-outcome Status-Of the six offenses
listed in Table 3.2.25 the number of juveniles
involved increasingly decreased in number.
The proportion of cases going to trial rather
than plea agreement was lower for Offenses 1
and 2. The pattern was reversed for Offense 3
onwards, the proportion of cases going to trial,
rather than plea agreement became close to 60
percent.

Outcome of Offense-Table 3.2.26 gives a
breakdown of the outcome of cases that went
through the JIB and DHS systems for the six
offenses being tracked. This was the final proc-
essing of juvenile delinquents and represented
the final sentencing stage. For the first and sec-
ond the most common form of treatment is
community service ranging between 21.3 to
31.0 percent of the cases.

For cases ranging three to six offenses, how-
ever, the treatment in the majority of instances
changed to sentencing for a period at the Youth
Rehabilitation Center (YRC)-between 31.3 to
39.5 percent of cases. It should be noted that







PRESENTATION ANALYSIS


Table 3.2.25 Pre-Outcome Status


Pre-Outcome Offense 1 Offense 2 Offense 3
Status Number Percent Number Percent Number Percent
Total 366 100.0 130 100 70 100
Trial 170 46.4 62 47.7 41 58.6
Plea Agreement 196 53.6 68 52.3 29 41.4

Table 3.2.25 Pre-Outcome Status: Continued

Pre-Outcome Offense 4 Offense 5 Offense 6
Status Number Percent Number Percent Number Percent
Total 36 100 26 100 25 100
Trial 21 58.3 15 57.7 16 64.0
Plea Agreement 15 41.7 11 42.3 9 36.0


for this population of juvenile delinquents for
whom the literature indicates there is high
marijuana usage, drug rehabilitation is listed as
a treatment only in Offenses 1, 2, 3 and 4 and
that at an extremely low rate.

3.2.3 Predicting Recidivism with the
St Thomas Department of Human
Services Data

This discussion focuses on forecasting the
chances of recidivism, or the chances that a ju-
venile will be arrested again after committing a
first offense. While most of the features of this
section are similar to those of 3.1.3.1-
particularly the approach, the methodology and
the model-there are two primary differences.
First, the data set is confined to only those ju-
veniles whose cases are on file at DHS for the
period 1987 to 1997. The number of observa-
tions in this data set-701-is less than half of
that of JIB. Second, in addition to the variables
employed in the model in section 3.1.3, there is
an additional variable that was recorded in
DHS case files, and that is the response to the
question, Has ... felt like a failure or ,, ill-
less?


3.2.3.1 Estimates of the St Thomas Depart-
ment of Human Services Data

The number of cases on which the DHS St
Thomas analysis is based one is 686 out of a
total of 701. Table 3.2.3.1 shows the results of
the estimation of the logistic regression model.
The likelihood ratio chi-square reported has a
probability value less than 0.01, indicating that
at least one of the coefficients for the explana-
tory variables is not 0. A glance at the values
of the parameter estimates of the explanatory
variables in column (1) are not largely different
from those in Table 3.1.3.1, because in many
respects they are measuring characteristics of
some of the same juvenile offenders.

The standard errors that are associated with the
parameter estimates are shown in column (2).
Column (3) shows that the variables Age, Age
squared and Sex are statistically significant. For
example, the chances that they are not signifi-
cant when one treats them as such are less than
1 in 10,000. The level of significance of US
Virgin Island born category hovers around
0.05, with Puerto Rico or Dominican Republic
Born category showing no statistical signifi-
cance. The relatively small size of the number









PRESENTATION ANALYSIS


Table 3.2.26 Outcome Status


Offense 1 Offense 2 Offense 3
Outcome Status Number Percent Number Percent Numberl Percent
Total 632 100.0 211 100.0 112 100.0
Crisis Stabilization Center 10 1.6 2 0.9 0 0.0
Youth Rehabilitation Center 62 9.8 38 18.0 35 31.3
Boys/Girls Group Home 0 0.0 1 0.5 0 0.0
Transfer to Correctional Facility 1 0.2 2 0.9 5 4.5
Drug Rehabilitation Center 10 1.6 7 3.3 1 0.9
Psychological evaluation 18 2.8 5 2.4 2 1.8
Dismised 154 24.4 30 14.2 10 8.9
Left jurisdiction 8 1.3 1 0.5 0 0.0
Case not prosecuted 37 5.9 7 3.3 3 2.7
Case pending 6 0.9 6 2.8 0 0.0
Community service 196 31.0 45 21.3 27 24.1
Counseling 25 4.0 12 5.7 5 4.5
restit = $2,000 8 1.3 1 0.5 1 0.9
Attend school regularly 1 0.2 0 0.0 0 0.0
Academic prog 0 0.0 0 0.0 1 0.9
read/subm bk rep 0 0.0 1 0.5 0 0.0
curf/prnt/princ 10 1.6 6 2.8 3 2.7
rpmnd/prin/j off 11 1.7 3 1.4 2 1.8
other 75 11.9 44 20.9 17 15.2

Table 3.2.26 Outcome Status: Continued

I Offense 4 Offense 5 Offense 6
Outcome Status Number Percent Numberl Percent Numberl Percent
Total 61 100.0 43 100.0 38 100.0
Crisis Stabilization Center 1 1.6 1 2.3 0 0.0
Youth Rehabilitation Center 22 36.1 17 39.5 15 39.5
Boys/Girls Group Home 0 0.0 0 0.0 0 0.0
Transfer to Correctional Facility 2 3.3 0 0.0 3 7.9
Drug Rehabilitation Center 1 1.6 0 0.0 0 0.0
Psychological evaluation 2 3.3 1 2.3 2 5.3
Dismised 9 14.8 5 11.6 3 7.9
Left jurisdiction 0 0.0 0 0.0 2 5.3
Case not prosecuted 2 3.3 1 2.3 0 0.0
Case pending 1 1.6 0 0.0 0 0.0
Community service 10 16.4 12 27.9 6 15.8
Counseling 0 0.0 1 2.3 2 5.3
restit = $2,000 0 0.0 0 0.0 0 0.0
Attend school regularly 0 0.0 0 0.0 0 0.0
Academic prog 0 0.0 0 0.0 0 0.0
read/subm bk rep 0 0.0 0 0.0 0 0.0
curf/prnt/princ 1 1.6 2 4.7 1 2.6
rpmnd/prin/j off 1 1.6 0 0.0 0 0.0
other 9 14.8 3 7.0 4 10.5







PRESENTATION ANALYSIS


Table 3.2.3.1 Parameter Estimates in Logistic Regression Predicting Recidivism:
St Thomas, Department of Human Services Data

Parameter Standard Pr> Adjusted
Variable Estimate Error Chi Square Odds Ratio
(1) (2) (3) (4)
Intercept -5.1437 2.2758 <0.0001

Age 1.2042 0.3443 <0.0001 3.334

Age Squared -0.5880 0.0130 <0.0001 0.943

Sex (1=Male) 0.4449 0.1154 <0.0001 2.435

Ethnicity
USVI Borna 0.5192 0.1903 0.0493 2.794

Puerto Rico or Dominican -0.6105 0.3819 0.9896 0.903
Republic Borna

Eastern Caribbean Borna 0.5995 0.2274 0.0554 3.027

Guardian (1=Mother and 0.0389 0.1421 0.0058 1.081
Stepfather or Foster Parents)b

Self-Esteem 0.3502 0.1890 <0.0001 2.015


Likelihood Ratio Chi Square = 141.2858 df= 8


Pr > Chi Square < 0.0001


a Compared to US Born.
b Compared to living with mother and father, mother only, father only, father and step mother, step
mother only, step father only, foster parents, or relatives.
Notes: N=686. Odds ratios greater than 1.0 indicate greater odds of recidivism.


of cases for this variable probably explains its
low level of significance. Guardian is statisti-
cally significant, and so is self-esteem, with
both levels of significance well below 0.01.

As was mentioned above in section 3.1.3, a
positive parameter estimate produces an odds
ratio greater than 1.0, and a negative odd ratio
results in an odds ratio below 1.0. The ad-
justed odds ratio for Age of 3.334 in column
(4) leads one to the same conclusion as that in
Table 3.1.3.1, that is, with a one-year increase
in age, the chances of committing at least one
other offense and being arrested, are more than
three times as likely as not committing another
offense, all other things being equal.


The odds ratio for Sex of 2.435 indicates that
males are about 2.5 times more likely than fe-
males to commit at least one other offense for
which they will be arrested. It is notable that
this ratio is somewhat higher than that for
males in the JIB data set. The odds ratios for
the Eastern Caribbean Ethnicity variable-
3.027-suggest that offenders from the East-
ern Caribbean are slightly more likely to be
repeat offenders than VI-born offenders-
2.794. Both of these groups are about three
times more likely to be repeat offenders than
US-born offenders.

The variable for Guardian does not appear to
exert as much influence in the DHS data set as
evidenced by the odds ratio of 1.081 compared








PRESENTATION ANALYSIS


Table 3.2.3.2 Predicted Probabilities in Logistic Regression: St Thomas, Department of Human
Services Data


Self- Proba-
Coefficients: a AgeFO AgeSq Sex Ethnicityl Ethnicity2 Ethnicity3 Guardian Esteem Odds ability


I (1) (2) (3) (4) (5) (6) (7) (8) (9) 1(10) (11)
Parameter
Estimates: -5.1437 1.2042 -0.0588 0.4449 0.5192 -0.6105 0.5995 0.0389 0.3502


Juvenile Profiles:


B1
B2
B3
B4
B5
B6
B7
B8

B9
B10

B11
B12
B13
B14
B15
B16
B17
B18
B19
B20
B21
B22


15 225 1 1 0 0 1

16 256 1 1 0 0 1


14 196 0 0 0 1 1
15 225 0 1 0 0 1


1 8.004
1 9.821
1 10.713
1 10.391
1 8.959
1 6.868
1 4.681
1 2.836

1 1.528
1 0.732
0 1.922
1 5.742
1 2.895
1 6.222
1 4.402
1 3.251
1 1.818
1 0.587
1 0.916
0 1.922
0 1.335
0 4.653


0.889
0.908
0.915
0.912
0.900
0.873
0.824
0.739

0.604
0.423
0.658
0.852
0.743
0.862
0.815
0.765
0.645
0.370
0.478
0.658
0.572
0.823


Variables:
AgeFO = Age of juvenile at first offense.
AgeSq = The square of the age of the juvenile at first offense.
Sex = Sex of the juvenile (1 = male, 0 = female).
Ethnicityl = Virgin Islander.
Ethnicity2 = Hispanic.
Ethnicity3 = Eastern Caribbean-born.
Ethnicity4 = US Mainland-born.
Guardian = Person with whom living (1 = mother and stepfather or foster parents, 0 = other guardian).
Self-esteem = Self-esteem of the juvenile, coded 1 if person does not feel like a failure, otherwise 0.







PRESENTATION ANALYSIS


to 1.897 in the JIB data. The odds ratio for
Self-esteem of 2.015, a variable not included in
the JIB data, indicates clearly that first-time
offenders who report that they experience low
esteem or feel worthless are twice as likely to
be a recidivist as an offender who expresses
self-assurance and feels self-confident.

3.2.3.2 The Prediction of Recidivism

As in Table 3.1.3.2, Table 3.2.3.2 show several
combinations of characteristics of juveniles that
might help to determine whether a juvenile will
be a recidivist or not. The set of variables is
the same as that in the former table with the
addition of the Self-Esteem variable. In a simi-
lar way, the set of juvenile profiles is repeated
so as to allow comparison with the JIB data.


Table 3.2.3.2 illustrates that because the pa-
rameter estimates of the coefficients in the
DHS data are in general close to those of the
JIB data, the odds and the associated probabili-
ties are not very different. Hence, in the first
10 profiles of juveniles between age eight and
age 17, male US Virgin Islands-born first-time
offenders living with a mother and stepfather or
foster parents and having low self-esteem are
predicted to be repeat offenders. This combi-
nation of demographic and social characteris-
tics predicts recidivists for all ages consistently
up to age 16. It is not until the youthful of-
fender with these attributes reaches age 17 that
the chances of recidivism are below 0.5, hence
predicted to remain a one-time offender. (See
section 3.1.3.6 for an explanation of interpreta-
tion of the prediction probabilities).


Figure 3.2.3.1 Predicted Chances of Recidivism: St Thomas DHS
For a USVI-Born Male Living With his Mother and a Stepson or Foster Parents

0.90


0.75








0.30


0.15
S0.600- --------------





0.00
B1 B2 B3 B4 B5 B6 B7 B8 B9 B10
Note: See text for profile characteristics Juvenile Profile







PRESENTATION ANALYSIS


St Croix Data: Description and Analysis

3.3 Youth Investigation Bureau Data
Variables

The initial assumption of the researchers was
that the same data collection instrument-the
Contact Card and Arrest Report-that was
utilized with the Investigation Bureau in St
Thomas could also be used with the St Croix
Youth Investigation Bureau data. This was not
entirely the case as the St Croix office deemed
it valuable to consider an additional variable for
which data were immediately available. The
data variables were as follows:
Sex
Age
Date of birth
Place of birth
Race/ethnic origin
Person with whom living
Birthplace of mother and father
School enrollment status
Education (highest level)
Employment status of guardian
Date of crime/incident
Type of crime/incident
Type of weapon used
Relative charged with criminal offense
Degree of criminal involvement of relative

The number of cases that occurred before 1987
were expected to be small since they repre-
sented only those juveniles that committed of-
fenses before the period of study but were still
active after 1987. Data between 1987 and 1997,
on the other hand, should have reflected all
cases that committed first offenses during those
years.

It appears very likely that the number of cases
used in data gathering particularly at YIB on
St Croix, did not constitute the entire popula-
tion of cases for the years stipulated and as an-
ticipated in the parameters of the study. This is
evident especially when observing Year of First


Offense data. This would mean that all the YIB
data would be affected by this absence of a full
quota of cases even when it is not immediately
obvious in the results. It is not possible to esti-
mate or devise the extent to which the data pre-
sented here do not represent the complete data
set. The actual data gathering process, though
guided by ECC, was carried out by YIB em-
ployees in order to comply with confidentiality
agreements.

There are concerns as to whether this data set
represents enough of the total population and
what possible effect the inclusion of additional,
but absent cases, would have had on the out-
come of variables. Thought should be given
also to how these factors restrict or facilitate
the drawing of conclusions using this data.

Following is an attempt to describe the data
that are presented in frequency tables, followed
by a section of cross tabulation and analysis
based on the items outlined in the data gather-
ing forms. Data are set out first for YIB, fol-
lowed by those for DHS. An attempt is made to
provide background information and brief de-
scriptions toward a clearer appreciation of the
data.


3.3.1 St Croix Data Description:
Investigation Bureau


Youth


Year of First Offense-The period from
which cases were chosen for this study was
1987 to 1997. All cases that were active during
that time should have been included during the
data gathering process. In some cases first and
subsequent offenses had been committed before
1987 and these years and offenses were in-
cluded in the data set as well.

The data in Table 3.3.1 suggest a very uneven
pattern in entry of juveniles as first time of-
fenders into the YIB system. These data sug-
gest that in 1989 and the period before had few
recorded juvenile first offenses-a total of








PRESENTATION ANALYSIS


Table 3.3.1 Year of First Offense

Year of Offense Juvenile Percent


Total


Before 1980
1980 to 1988
1989
1990
1991


1992
1993
1994
1995
1996


1821 100.0
1 0.1
1 0.1
1 0.1
102 5.6
435 23.9
361 19.8
18 1.0
13 0.7
753 41.4
136 7.5


three during the period before 1980 and in
1989-that required intervention by the YIB
system. This pattern appears extremely unusual
and might likely be a reflection of administra-
tive recordkeeping up to the time of data gath-
ering, and not necessarily an accurate account
of the cases that were recorded by the YIB sys-
tem during those years.

The cluster of years beginning in 1990 and ex-
tending to 1996 show juveniles as relatively
active in committing first offenses. During that
period, however, the data suggest a steep dip of
two years where the percent of offenders fell to
one and below-(1993, 1994). This is followed
by a period of increased activity that peaked at
41.1 percent in 1995, followed by a drop to 7.5
percent in 1996. Once again, it is both an inter-
esting and unusual phenomenon that first-time
delinquent behavior would drop drastically to
as low as zero in 1997 after such an intense pe-
riod of activity, and suggests that this might be
a reflection of a deficiency in administrative
recordkeeping. No serious significance should
therefore be placed on the distribution of first
offenses over time using this data set.


Type of First Offense-Table 3.3.2 lists first-
time offenses under very broad categories. Part
1 Felony accounted for the largest single of-
fense type-28.7 percent-and included crimes
such as robbery, felonious assault, burglary,
and grand larceny. Part II Felony-11.8 per-
cent-covered crimes such as possession of a
controlled substance, possession of unlicensed
firearm, possession of stolen property, and un-
authorized use of vehicle. A large proportion-
41.0 percent-of offense types fell under the
category other, compiled from a number of un-
defined offenses.

Weapon Used in First Offense-According to
Table 3.3.3, the highest proportion of cases fell
in the other category, implying a number of un-
defined objects that may be deemed weapons
were used by juveniles while carrying out de-

Table 3.3.2 Type of First Offense

Type of Offense Juvenile Percent
Total 1824 100.0
Part I Felony 523 28.7
Part II Felony 215 11.8
Misdemeanor 187 10.3
Incidence 152 8.3
Other 747 41.0

Table 3.3.3 Weapon Used in First
Offense

Weapon I Juvenile Percent
Total 634 100.0
Gun 101 15.9
Knife 50 7.9
Hand 187 29.5
Bottle 8 1.3
Penis 41 6.5
None 8 1.3
Other 239 37.7








PRESENTATION ANALYSIS


linquent activities. Hand is the next highest
category-29.5 percent-suggesting some
amount of unplanned, spontaneous involve-
ment. This is followed by gun and knife-15.9
and 7.9 percent, respectively. Each suggests
some level of deliberate and planned armed ac-
tivity. Penis-6.5 percent-implies the com-
mitting of sexual offenses.

Total Number of Offenses-Table 3.3.4 indi-
cates that the vast majority of juveniles-84.6
percent-entered the system by committing a
first offense and exited the system, never again
committing an offense that led to arrest as a
juvenile. Repeat offenders that committed two
offenses before exiting the system made up 10
percent of the cases. The percent of juveniles
committing three or more offenses was rela-
tively small-5.4 percent.

3.3.2 St Croix: Youth Investigation Bureau
Data Relationships

Age at First Offense by Sex-Table 3.3.5 in-
dicates juvenile delinquency levels were very
low-below 1 percent-between ages 7 to 9
and increased to 2 percent for 10-year-olds. De-


Table 3.3.4 Total Number of Offenses

No. Of Offenses Juvenile Percent
Total 1824 100.0
1 1544 84.6
2 183 10.0
3 49 2.7
4 18 1.0
5 4 0.2
6 17 0.9
7 3 0.2
8 2 0.1
9 0 0.0
10 3 0.2
11 0 0.0
12 1 0.1


linquency rates continued to increase with age,
climbing to 10.3 percent for 13-year-olds and
remaining high until age 17. The peak of activ-
ity occurred at age 15, with 22.3 percent of ju-
veniles committing their first offense at this
age.


Table 3.3.5 Age at First Offense by Sex

Age at First Total Male Female
Offense Number Percent Number Percent NumberlPercent
Total 1797 100.0 1285 100.0 511 100.0
Age 7 10 0.6 6 0.5 4 0.8
Age 8 12 0.7 9 0.7 3 0.6
Age 9 15 0.8 10 0.8 5 1.0
Age 10 36 2.0 33 2.6 3 0.6
Age 11 52 2.9 35 2.7 17 3.3
Age 12 83 4.6 52 4.0 30 5.9
Age 13 185 10.3 123 9.6 62 12.1
Age 14 324 18.0 205 16.0 119 23.3
Age 15 400 22.3 282 21.9 118 23.1
Age 16 368 20.5 263 20.5 105 20.5
Age 17 312 17.4 267 20.8 45 8.8







PRESENTATION ANALYSIS


When the data are broken down by gender, it is
observed that a pattern of increased activity oc-
curred for both sexes between ages 13 to 17.
The largest percent of males committed their
first offense between ages 15 and 17, while the
corresponding period for females was between
ages 14 to 16. By age 17, females had dropped
to 8.8 percent of cases having a first offense,
while the males were still relatively active, with
20.8 percent committing a first-time offense.
These data suggest that females became in-
volved in delinquent activity at a younger age
than males. Females also abandoned their de-
linquent activity at an earlier age than did
males. This observation coincides with other
child development concepts that assign higher
levels of physical and social development to
females at an earlier age than to males. How-
ever, given the uncertainty surrounding the
completeness of the data set being used here, it
would be inadvisable to apply such conclusions
on any broad scale. Additional data could
change the outcome of these variables.

Juvenile's Place of Birth-The data in Table
3.3.6 permit a comparison between the percent-
age of juvenile offenders by ethnic origin and
the actual percentage in the population at large.
The data show that the percentage of juvenile
offenders-7-to-17-year olds-born in the US
Virgin Islands was 79.9 percent. This figure
compares with 75.8 percent that made up the 5-
to-19-year olds that were counted in the 1990
Census of Population. These data suggest, pro-
portionately, that slightly more Virgin Islanders
were among the offenders than their numbers in
the population would suggest. The percentage
of offenders born in Puerto Rico, Antigua and
Barbuda and St Lucia were also slightly higher
than their percentage in the total population
would suggest.

Juvenile's Place of Birth by Sex-Table 3.3.7
shows males and females having similar per-
centages of juvenile delinquents based on the
total delinquents for each category. Here again,


Table 3.3.6 Juvenile's Place of Birth

Place of Birth Number Percent
Total 1825 100.0
US Virgin Islands 1458 79.9
Puerto Rico 48 2.6
United States 129 7.1
Antigua or Barbuda 36 2.0
Dominica 22 1.2
St Kitts or Nevis 32 1.8
St Lucia 42 2.3
Other Eastern Caribbean 39 2.1
Other 19 1.0

for both males and females, persons born in the
US Virgin Islands constituted by far the largest
proportion of delinquents-79.7 and 80.4 per-
cent respectively.

The next highest category was made up of
those born in the US-6.6 percent males and
8.3 percent females. The remainder of the juve-
nile delinquents born in other Caribbean islands
or group of islands ranged between 1.1 and 2.6
percent, males and 1.0 and 3.3 percent, fe-
males.

Age at First Offense by Juvenile's Place of
Birth-In Table 3.3.8 suggests that juveniles
from all ethnic groups under the age of 10 had
a relatively low rate of committing their first
offense, compared to the other age categories.
Significant increases in the rate occurred be-
tween ages 10 to 13 for all ethnic groups but
with a noticeable difference in percentage for
those born in Puerto Rico where it rose to 29.5
percent, which was a much higher rate than for
similarly aged juveniles who were represented
in the other ethnic categories. The peak age
group for which first-time delinquent activity
across all ethnic groups was between 14 to 16
year-olds with a falling off at the over 16 age
category. Juveniles of Puerto Rican or Domini-







PRESENTATION ANALYSIS


Table 3.3.7 Juvenile's Place of Birth by Sex

Total Male Female
Place of Birth Number Percent Number Percent Number Percent

Total 1825 100.0 1303 100.0 521 100.0
US Virgin Islands 1458 79.9 1038 79.7 419 80.4
Puerto Rico 48 2.6 31 2.4 17 3.3
United States 129 7.1 86 6.6 43 8.3
Antigua or Barbuda 36 2.0 28 2.1 8 1.5
Dominica 22 1.2 17 1.3 5 1.0
St Kitts or Nevis 32 1.8 26 2.0 6 1.2
St Lucia 42 2.3 34 2.6 8 1.5
Eastern Caribbean 39 2.1 29 2.2 10 1.9
Other 19 1.0 14 1.1 5 1.0


can Republic descent were especially likely to
not have committed their first offense when
over 16 years old-8.2 percent-when com-
pared to those juveniles in the other place of
birth categories-16.4 to 50.0 percent.

Age at First Offense by Mother's Place of
Birth-In Table 3.3.9, a similar pattern appears
as in Table 3.3.8 above. First-time delinquent
activity remained low for juveniles under 10
years across all ethic groups, but increased sig-
nificantly among the 10 to 13 age groups.
There was a peaking of delinquent activity for
juveniles with in the 14 to 16 age category for
all ethnic groups.


The trend that seems to follow juvenile delin-
quents born in Puerto Rico was apparent here
in the case of juveniles with mothers of Puerto
Rican or Dominican Republic descent. For in-
stance, juveniles with mothers who were born
in Puerto Rico or Dominican Republic had a
higher rate of a first offense in the 10 to 13 age
group-30.2 percent-than juveniles with
mothers born elsewhere-16.7 to 16.9 percent.
This would seem to suggest that a greater pro-
portion of juveniles with a mother from a
Puerto Rican or Dominican Republic back-
ground become involved in delinquent activity
between ages 10 to 13. The relative fall off in
first time offenses by juveniles over age 16 was


Table 3.3.8 Age at First Offense by Juvenile's Place of Birth

Place of Birth
Puerto Rico or
Age at First Virgin Islands Dominican Republic United States Eastern Caribbean Other
Offense Number Percent Number Percent Number Percent Number Percent Number Percent
Total 1431 100.0 61 100.0 128 100.0 169 100.0 6 100.0
Under 10 yrs 30 2.1 2 3.3 4 3.1 1 0.6 0 0.0
10to 13yrs 286 20.0 18 29.5 24 18.8 26 15.4 1 16.7
14 to 16 yrs 863 60.3 36 59.0 79 61.7 111 65.7 2 33.3
Over 16 yrs 252 17.6 5 8.2 21 16.4 31 18.3 3 50.0







PRESENTATION ANALYSIS


Table 3.3.9 Age at First Offense by Mother's Place of Birth

Mother's Place of Birth
Puerto Rico or
Age at First Virgin Islands Dominican Republi United States Eastern Caribbean Other
Offense Number Percent Number Percent Number Percent Number Percent Number Percent
Total 1355 100.0 86 100.0 90 100.0 219 100.0 6 100.0
Under 10 yrs 26 1.9 2 2.3 2 2.2 4 1.8 0 0.0
10 to 13 yrs 268 19.8 26 30.2 15 16.7 37 16.9 1 16.7
14 to 16 yrs 823 60.7 49 57.0 61 67.8 136 62.1 2 33.3
Over 16 yrs 238 17.6 9 10.5 12 13.3 42 19.2 3 50.0

especially apparent for those juveniles with Place of Birth. The involvement in delinquent
mothers of Puerto Rican or Dominican Repub- activity became apparent in the 10 to 13 age
lic descent-10.5 percent. In contrast, juve- group for all juveniles across all place of birth
niles beyond the age of 16 with mothers who categories. The rate of first time offenses also
were not born in Puerto Rico or the Dominican peaked at the 14 to 16 age category for all juve-
Republic had rates of a first-time offense that niles, regardless of their father's place of birth.
ranged from 13.3 percent to 50 percent.
Juvenile's Race by Sex-Table 3.3.11 shows
Age at First Offense by Father's Place of that of the total number of juvenile delinquents
Birth-It is of interest to note that the pattern who committed first offenses, 76.8 percent
apparent in the total population of juvenile de- were black, 1.7 percent white, a small percent
linquents and those with fathers born in Puerto was Asian, and an undefined Other was 21.4
Rico or Dominican Republic in Table 3.3.10 is percent. These proportions were generally con-
similar to that found in two previous tables de- sistent for males and females. Slightly more
picting Age at First Offense by Place of Birth males than females-77.4 percent versus 75.1
Group, and Age at First Offense by Mother's percent-were among black offenders.

Table 3.3.10 Age at First Offense by Father's Pace of Birth

Father's Place of Birth
Puerto Rico or

Age at First Virgin Islands Dominican Republic United States Eastern Caribbean Other
Offense Number Percent Number Percent Number Percent Number Percent Number Percent

Total 728 100.0 88 100.0 57 100.0 187 100.0 4 100.0
Under 10 yrs 15 2.1 3 3.4 1 1.8 3 1.6 0 0.0
10to 13yrs 142 19.5 24 27.3 12 21.1 28 15.0 1 25.0
14 to 16 yrs 445 61.1 55 62.5 37 64.9 118 63.1 0 0.0
Over 16 yrs 126 17.3 6 6.8 7 12.3 38 20.3 3 75.0







PRESENTATION ANALYSIS


Table 3.3.11 Juvenile's Race by Sex

Total Male Female
Race Numberl Percent Number Percent Numberl Percent
Total 1819 100.0 1299 100.0 519 100.0
Black 1397 76.8 1006 77.4 390 75.1
White 31 1.7 22 1.7 9 1.7
Asian 1 0.1 1 0.1 0 0.0
Other 390 21.4 270 20.8 120 23.1


Juvenile's Guardian by Sex of Juvenile-The
data in Table 3.3.12 suggests that the majority
of juvenile offenders-78.7 percent-lived
with their mother; the second highest percent-
18.3 percent-lived with their mother and fa-
ther. The remainder of the guardian categories
ranged between 0 to 1.3 percent.

When looked at by gender, both male and fe-
male categories of juvenile delinquents fol-
lowed similar patterns within similar ranges.
Female offenders were slightly more like to
live with both their mother and father than were
male offenders-21.3 percent and 17.0 percent,


respectively. All the remaining of guardian
categories experienced relatively low propor-
tions of the total number of juvenile delin-
quents. One can assume with some confidence
that this is quite likely a reflection of the repre-
sentation of these groups in the general Virgin
Islands' population, namely, the majority of
children in this age group live with Mothers
alone and the second largest group with Mother
and Father. However, a categorical statement
cannot be made on this matter since there is no
corresponding data available for comparison in
the general population data on St Croix.


Table 3.3.12 Juvenile's Guardian by Sex

Total Male Female
Guardian Number Percent Numberl Percent Numberl Percent

Total 1726 100.0 1236 100.0 489 100.0
Mother and Father 315 18.3 210 17.0 104 21.3
Mother 1359 78.7 989 80.0 370 75.7
Father 12 0.7 10 0.8 2 0.4
Mother and Step Father 3 0.2 1 0.1 2 0.4
Father and Step Mother 1 0.1 0 0.0 1 0.2
Step Mother 0 0.0 0 0.0 0 0.0
Step Father 0 0.0 0 0.0 0 0.0
Foster Parents 1 0.1 0 0.0 1 0.2
Relatives 23 1.3 19 1.5 4 0.8
Other 12 0.7 7 0.6 5 1.0







PRESENTATION ANALYSIS


Table 3.3.13 School Enrollment

Status Number Percent
Total 1789 100.0
Enrolled 1722 96.3
Not Enrolled 67 3.7


School Enrollment at Time of Arrest-Table
3.3.13 indicates that only a small proportion of
juvenile delinquents-3.7 percent-were not
enrolled in school at the time of arrest. This
suggests that being enrolled in school is not a
deterrent to youngsters committing offenses. A
more comprehensive picture might be gained if
data were available for analysis on levels of
school and class attendance, and even achieve-
ment measures.

School Type at Time of Arrest-Table 3.3.14
shows data on school type at which juvenile
delinquents were enrolled at the time of first
offense. These data seem to correlate with pre-
vious sets of data that reflected age and first
offense and show that a relatively low percent-


age of juvenile delinquents committed their
first offense in elementary school-5.5 per-
cent-increasing to 23.3 percent in Jr. high
school and rising to 71.1 percent by high
school.

Mother's Place of Birth by Juvenile's Sex-
Table 3.3.15 reflects a similar pattern encoun-
tered in juvenile's place of birth. The majority
of juvenile delinquents, 77.3 percent, came
from mothers born in the US Virgin Islands.
The second highest proportion was from the
United States-5.1 percent. Puerto Rico and St
Lucia each had around four percent. The pro-
portion of juvenile delinquents with mothers
born in other Caribbean islands or island group
fell to approximately between one and three
percent.

Table 3.3.14 School Type

School Number Percent
Total 1684 100.0
Elementary 93 5.5
Jr. High 393 23.3
High 1198 71.1


Table 3.3.15 Mother's Place of Birth by Sex of Juvenile

Total Male Female
Place of Birth Number Percent Number Percent Number Percent

Total 1785 100.0 1277 100.0 507 100.0
Virgin Islands 1380 77.3 986 77.2 394 77.7
Puerto Rico 72 4.0 48 3.8 24 4.7
United States 91 5.1 56 4.4 34 6.7
Antigua or Barbuda 38 2.1 28 2.2 10 2.0
Dominica 27 1.5 21 1.6 6 1.2
St Kitts or Nevis 36 2.0 28 2.2 8 1.6
St Lucia 74 4.1 58 4.5 16 3.2
Eastern Caribbean 46 2.6 36 2.8 10 2.0
Other 21 1.2 16 1.3 5 1.0







PRESENTATION ANALYSIS


Table 3.3.16 Mother's Employment by Sex of Juvenile

Employment Total Male Female
Status Number Percent Number Percent Number Percent
Total 1398 100.0 1000 100.0 397 100.0
Employed 1321 94.5 939 93.9 381 96.0
Unemployed 77 5.5 61 6.1 16 4.0


Mother's Employment by Sex of Juvenile-
Table 3.3.16 shows that only a relatively small
proportion, 5.5 percent, of the total juvenile
delinquents were of mothers who were unem-
ployed. The percentage of female offenders
with mothers who were employed was slightly
higher than it was for male offenders. This may
suggest that employed mothers were away from
the home for extended hours especially at times
when children returned home from school. Data
do not indicate if any of those mothers held
more than one job or worked overtime, factors
that would obviously have added to the mother-
child separation and left unsupervised children
exposed and vulnerable to involvement in de-
linquent behavior.


Father's Place of Birth by Juvenile's Sex-
Table 3.3.17 indicates that 68.8 percent of juve-
nile offenders had fathers who were born in the
Virgin Islands; those born of Puerto Rican fa-
thers made up 7.2 percent and offenders with
St Lucian fathers made up 6.7 percent. Female
offenders were slightly more likely to have fa-
thers who were born in the US Virgin Islands
and Puerto Rico than male offenders.

Father's Employment by Juvenile's Sex-
The employment statistics of fathers of juvenile
offenders at time of first offense follows
closely that of mother's employment statistics
(Table 3.3.18). A small percentage of fathers-
5.4 percent-were unemployed. Unemploy-


Table 3.3.17 Father's Place of Birth by Sex of Juvenile

Total Male Female
Place of Birth Number Percent Number Percent Number Percent

Total 1079 100.0 757 100.0 321 100.0
Virgin Islands 742 68.8 508 67.1 233 72.6
Puerto Rico 78 7.2 50 6.6 28 8.7
United States 57 5.3 43 5.7 14 4.4
Antigua or Barbuda 28 2.6 21 2.8 7 2.2
Dominica 17 1.6 12 1.6 5 1.6
St Kitts or Nevis 37 3.4 29 3.8 8 2.5
St Lucia 72 6.7 59 7.8 13 4.0
Eastern Caribbean 34 3.2 25 3.3 9 2.8
Other 14 1.3 10 1.3 4 1.2







PRESENTATION ANALYSIS


Table 3.3.18 Father's Employment by Sex of Juvenile

Employment Total Male Female
Status Number Percent Number Percent Number Percent
Total 719 100.0 517 100.0 201 100.0
Employed 680 94.6 488 94.4 191 95.0
Unemployed 39 5.4 29 5.6 10 5.0


ment was slightly higher for the fathers of male
juvenile offenders-5.6 percent compared to
the fathers of female offenders-5.0 percent.

Relationship Between Relatives With Previ-
ous Criminal Offenses and Recidivism-
Officers of the YIB in St Croix expressed the
view that, based on their experience, there was
an observable link between the number of re-
peat offenders and homes with relatives that
had previously been charged with criminal of-
fenses. The data in Table 3.3.19 were compiled
to investigate this assumed relationship.

From the data in the table, it can be observed
that of the total 1,782 youth charged with of-
fenses, 505 or 28.3 percent of them came from
homes in which relatives had been previously
charged with criminal offenses. The table
highlights that of the 505 homes with previ-
ously charged relatives, 35.4 percent of them
were recidivists compared to the 7.7 percent
recidivists that came from homes in which


there were no relatives charged with offenses.
The data in the table were further subjected to
more rigorous statistical analysis to test the
proposition of a significant relationship be-
tween households with relatives that had previ-
ously been charged and recidivism among
youth. The test does strongly support the view
that there is a very strong relationship between
the rate of recidivism and households with rela-
tives that have been charged with criminal ac-
tivities. The p-value at the bottom of the table
indicates that this conclusion is likely to be in-
correct in less than 1 in 10,000 times.

Juvenile's Age by Current or Highest
Grade-Table 3.3.20 suggests that juveniles
offenders tended to be in grade levels consis-
tent with their age. There were a few excep-
tions. For instance, there were several 14 and
15 year olds who had not reached Grade 8 yet
and some 16 and 17 year olds who were at a
grade level less than 10. The existence of
some juveniles being at a grade level much


Table 3.3.19 Recidivism and Relatives Charged With Offenses

Recidivism
Relatives Yes No Total
Charged Number Percent Number Percent Number Percent
Total 277 15.5 1 ,505 84.5 1 ,782 100.0
Yes 179 35.4 326 64.6 505 100.0
No 98 7.7 1,179 92.3 1,277 100.0
72 = 212.6 df= 1, p < 0.0031
Odds ratio = 6.61
5.02







PRESENTATION ANALYSIS


Table 3.3.20 Juvenile's Age by Current or Highest Grade

Age in Total Grade 1-6 Grade 7 Grade 8
years Number Percent Number Percent Number Percent Numberd Percent
Total 1670 100.0 133 100.0 112 100.0 237 100
7-9 36 2.2 33 24.8 0 0.0 1 0.4
10-12 165 9.9 95 71.4 58 51.8 7 3.0
13 178 10.7 2 1.5 50 44.6 115 48.5
14 311 18.6 2 1.5 2 1.8 101 42.6
15 379 22.7 0 0.0 2 1.8 12 5.1
16 347 20.8 0 0.0 0 0.0 0 0.0
17 254 15.2 1 0.8 0 0.0 1 0.4

Table 3.3.20 Juvenile's Age by Current or Highest Grade: Continued

Age in Grade 9 Grade 10 Grade 11 Grade 12
years Number Percent Number Percent Number Percent Numberl Percent
Total 358 100.0 450 100.0 311 100.0 69 100.0
7-9 0 0.0 0 0.0 2 0.6 0 0.0
10-12 4 1.1 0 0.0 1 0.3 0 0.0
13 8 2.2 3 0.7 0 0.0 0 0.0
14 196 54.7 7 1.6 0 0.0 3 4.3
15 134 37.4 228 50.7 3 1.0 0 0.0
16 14 3.9 203 45.1 129 41.5 1 1.4
17 2 0.6 9 2.0 176 56.6 65 94.2


lower than would be expected given their age
may indicate the propensity for such juveniles
to be underachievers. However, since social
promotion was practiced throughout the public
school system of the US Virgin Islands at the
period referenced by the study, it is difficult to
assess whether underachievement as indicated
by grade level was a factor influencing behav-
ior.

3.3.3 Predicting Recidivism with the Youth
Investigation Bureau Data

The methods and procedures utilized in this
section and similar to those that were applied to
the St Thomas JIB and DHS data above. (See
Sections 3.1.3.1, 3.1.3.2, and 3.1.3.4 for a thor-


ough discussion of the approach, methodology
and recidivism model.)

There is a notable difference between the data
of the Youth Investigation Bureau (YIB) in St
Croix and the JIB data in St Thomas. Some of
the variables that were significant in the
St Thomas model showed no statistically sig-
nificant association with recidivism, while a
new set of variables were clearly very good
predictors of repeat arrests. For example, it
was shown above empirically that age is a very
strong predictor of recidivism in St Thomas in
both the JIB and DHS data sets, but in the
St Croix data set, it is a very poor predictor.
Similarly, ethnicity and the role of the parents
or guardians fail to exert any statistically strong







PRESENTATION ANALYSIS


relationship with repeat offense behavior.

The immediate suggestion is that both sets of
juvenile come from significantly different
populations or social milieus, and that attempts
to prescribe similar explanations for self-same
behaviors could be enormously misleading.
The variables included in the St Croix models
reflect such differences.

3.3.3.1 Definition of Variables

Dependent Variable: As in the previous cases
for St Thomas, the outcome or dependent vari-
able of interest is recidivism. Each juvenile
was classified in one of two discrete categories.
If the youth were charged with more than one
offense, (s)he was classified as a repeat of-
fender or recidivist. The other classification
comprised youth who were charged with a sin-
gle offense. This latter group was made the
referent group.

Independent or Explanatory Variables: Ide-
ally, it would have been most valuable to com-
pare the impact of the same variables on St
Croix as were used for St Thomas, but the dif-
ferences in data precluded this. For example,
preliminary analysis showed considerable in-
consistency with the age distribution of the YIB
data, hence they were excluded from the analy-
sis. This is a significant loss since theory does
point to age as critical in this prediction model.
Additionally, one new variable that relates to
the involvement of relatives of the juvenile in
criminal activity appears to be an essential pre-
dictor.

i) Gender: As in the previous models, this
variable is dichotomous and refers to
the sex of the juvenile. Female is the
referent category.

ii) School Enrollment: This variable indi-
cates whether the juvenile was enrolled
in school at the time of the first offense.
The referent group was "not enrolled in


school at the time of first offense".

iii) Grade: This variable is dichotomous
and refers to whether the offender was
in Grade 12 or in a grade level less than
12 when first charged with an offense.
The referent group was Grade 12.

iv) Relative Charged: This dichotomous
variable classified each youth as to
whether (s)he had a relative who had
been charged with a crime previously.
The referent group was those who did
not have relatives that had been previ-
ously charged with a criminal offense.

3.3.3.2 Estimates of the St Croix Youth
Investigation Bureau Data

The total number of cases that were recorded in
the study period was 1,827. In some instances,
data items were missing and some cases had to
be dropped from a particular analysis. In the
application of the logistic regression model,
1,796 valid cases were used.

Four variables in the YIB data set show strong
association with repeat offenses. The sex of the
arrested youth remains a reliable predictor. En-
rollment in school at the time of the first arrest
is also a good indicator of the type of subse-
quent behavior that might be expected from
first-time arrested youth.

A particularly strong predictor is related to
whether the first-time arrested youth had rela-
tives that had been charged with criminal of-
fenses at any time previously. Data analysis in
3.2.2 (Table 3.3.19) illustrated that there is a
statistically significant relationship between
deviant behaviors in youth when a relative has
been known to have been charged with criminal
offenses in the past. The odds ratio as a meas-
ure of the strength of the relationship confirms
that a first-time offender with relatives charged
with criminal offenses is seven times more
likely to be a repeat offender than one who







PRESENTATION ANALYSIS


Table 3.3.3.1 Parameter Estimates in Logistic Regression Predicting Recidivism:
St Croix, Youth Investigation Bureau Data


Sex (1=Male) 0.3389 0.0937 0.0003 1.970


School Enrollment (1=Enrolled)

Grade (1=Less than Grade 12)


0.6398

0.4324


0.2513

0.2265


0.0109

0.0562


3.595

2.375


Relative Charged (1=Yes) 0.9664 0.0757 0.0001 6.908
Likelihood Ratio Chi Square = 204.1609 df = 4 Pr > Chi Square < 0.0001

Notes: N=1796. Odds ratios greater than 1.0 indicate greater odds of recidivism.


does not have relatives charged with such of-
fenses.

The data in Table 3.3.3.1 are interpreted simi-
larly to those in 3.1.3.1 and 3.2.3.1 above. The
odds ratio of 1.970 for Sex denotes that the pre-
dicted odds of recidivism for a first-time male
offender are about twice the odds for a female
offender or 97 percent higher than the odds for
a female first-time offender.

The adjusted odds ratio of a first-time offender
who is enrolled in school is 3.595. This means
that the odds of an offender who was not en-
rolled in school at the time of his or her first
arrest becoming a repeat offender are at least
three and one-half times the odds of someone
who is enrolled. Equivalently, it can be said
that the odds of repeat arrests of a juvenile who
was enrolled at school are only about 28 per-
cent of the odds of someone who was not en-
rolled in school.

The grade level of a juvenile is also an indica-
tor of what to anticipate in a youth's behavior.
The odds ratio for Grade in Table 3.3.3.1 is
2.375. This represents the predicted odds of


recidivism of a youth who was in a grade below
12 at the time of first arrest. The odds of re-
cidivism for youngsters who are below Grade
12 are more than twice-2.375-as high as
those who committed their first offense in
Grade 12. In other words, the odds of recidi-
vism for juveniles in Grade 12 are only 42 per-
cent of the odds of a youth who had not
reached Grade 12.

It is to be recalled that in section 3.3.3 it was
shown that there was a very strong association
between repeat offense behavior and the pres-
ence in the lives of these youth of relatives who
have been charged with criminal offenses
(Table 3.3.19). This relationship is maintained
at about the same level even in the presence of
other explanatory variables. The table shows
that the adjusted odds ratio for the Relatives
Charged variable is 6.908. This value asserts
that the predicted odds for repeat arrests from a
first-time male offender with relatives who
have been charged with criminal behavior in
the past are about seven times the odds of a
first-time offender without relatives having
such a criminal history.








PRESENTATION ANALYSIS


Table 3.3.3.2 Predicted Probabilities in Logistic Regression: St Croix, Youth Investigation Bureau Data


Coefficients: a Sex Enrollment Grade Rel. Charged Odds Probability
(1) (2) (3) (4) (5) (6) (7)
Parameter
Estimates: -1.5531 0.3389 0.6398 0.4324 0.9664
Juvenile Profiles:
C1 1 1 1 1 2.281 0.695
C2 0 1 1 1 1.625 0.619
C3 1 1 0 1 1.480 0.597
C4 1 0 1 1 1.203 0.546
C5 0 1 0 1 1.055 0.513
C6 1 1 1 0 0.868 0.465
C7 0 0 1 1 0.857 0.462
C8 1 1 0 0 0.563 0.360
C9 0 0 0 1 0.556 0.357
C10 1 0 1 0 0.458 0.314
C11 0 1 0 0 0.401 0.286
C12 0 0 0 0 0.212 0.175

Variables:
Sex = Sex of the juvenile (1 = male, 0 = female).
Enrollment = Coded 1 if not enrolled in school at first arrest, otherwise 0.
Grade = Coded 1 if school grade is less than 12th grade, otherwise 0.
Rel. Charged = Coded 1 if a relative was charged with a criminal offense, otherwise 0.


3.3.3.3 The Prediction of Recidivism

The probabilities in Table 3.3.3.2 indicate the
chances of recidivism or non-delinquent behav-
ior-using YIB St Croix data-for a given
combination of the variables under study. (See
also Figures 3.3.3.1 and 3.3.3.2). Again, for
odds-in column (6)-greater than 1.0 or for
probabilities-in column (7)-greater than 0.5,
the combination of characteristics predicts that
the juvenile will commit offenses for which he
or she will be arrested. And for odds less than
1.0 or probabilities less than 0.5, the prediction
is that a first-time offender will not be arrested
as a juvenile again.


The profile for juvenile Cl in the table is inter-
preted as follows. A first-time male offender
who is not enrolled in school at the time of his
arrest, who has not reached Grade 12, and who
has relatives that had been charged with crimi-
nal offenses has odds of 2.281 and probability
of 0.695, thus indicating that this youth has a
high probability of being a recidivist. Profile
C2 is that of a female who is enrolled in school
in Grade 12 with relatives charged for commit-
ting criminal offenses is predicted to be a one-
time offender and not be a recidivist. The
youth that is represented by profile C11 is that
of a female with one offense, who is enrolled in
Grade 12 at the time of arrest, and who does









PRESENTATION ANALYSIS


Figure 3.3.3.1 Predicted Changes of Recidivism: St Croix YIB
For Various Combinations of Youth Characteristics


0.75



0.60

8 -

S0.45
0
0

-E 0.30
.Q0


0.15



0.00 -------------
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12
Note: See text for profile characteristics Juvenile Profile


Figure 3.3.3.2 Age at First Offense: St Croix
Youth Investigation Bureau Data





20
S Male
S eFemale

15



10



5



0
7 8 9 10 11 12 13 14 15 16 17
Age







PRESENTATION ANALYSIS


not have relatives charged with any criminal
offense. A female offenders with these charac-
teristics is predicted to have a relatively low
chance of being a repeat offender. It is notable
that for most variable combinations that include
the 'Relatives Charged' variable, the prediction
is that the juvenile will be a repeat offender.
This is consistent with all of the foregoing
analyses that include this particular variable.

3.4 Department of Human Services Data
Variables

The following data items were collected from
case files at the St Croix Department of Human
Services (DHS):
Sex
Immigration Status
Age
Place of Birth
Race
Household Status of Parents
Age of Parents
Marital Status of Parents
Guardian
Marital Relationship of Natural Parents
Number of Siblings in the Household
School Enrollment
Employment Status Juvenile
Children of Juvenile
Psychological Assessment of Juvenile
Sexual Activity of Juvenile
Pre-outcome Status of Juvenile
Type of Offense
Year of Offense
Outcome of Offense
Biological Father Known by Juvenile

There were 888 records in the St Croix district
DHS data set. These records consist of all ju-
veniles who were on DHS files during the pe-
riod 1987 to 1997. As a result of missing data,
many of the tables below will reflect differing
totals.


3.4.1 St Croix Data Description: Department
of Human Services

Juvenile's Sex-According to Table 3.4.1 a
large proportion of the total juvenile offend-
ers-85.9 percent-were males compared to
14.1 percent who were females. The age cate-
gory here refers to seven-to 17-year-olds: as a
comparison, the five-to 19-year-olds in the total
St Croix population had a 49.9 to 50.1 male-
female ratio (1995 Population and Housing
Survey). This over representation of males in-
volved in delinquent behavior is a common and
well-known phenomenon, though the reasons
for the occurrence are still not as clear.

Immigration Status-In Table 3.4.2 US citi-
zens made up the overwhelming majority of
juvenile offenders-94.1 percent-whereas
juveniles of resident alien status accounted for
just 5.9 percent. The closest corresponding age
group in the St Croix general population for
which data are available is five to 19; US citi-
zens and resident aliens or permanent residents
in this group made up 88.4 and 10.8 percent,
respectively (1995 Population and Housing
Survey).

Table 3.4.1 Juvenile's Sex


Juvenile
Sex Number Percent

Total 885 100.0
Male 760 85.9
Female 125 14.1

Table 3.4.2 Immigration Status

Immigration Juvenile
Status Number Percent

Total 717 100.0
US Citizen 675 94.1
Resident Alien 42 5.9








PRESENTATION ANALYSIS


Table 3.4.3 Age of Juvenile

Juvenile
Age in years Number Percent
Total 860 100.0
7-9 3 0.3
10-12 31 3.6
13 57 6.6
14 151 17.6
15 202 23.5
16 219 25.5
17 197 22.9


Juvenile's Age-The data in Table 3.4.3, Juve-
niles' Age suggest that delinquent activity re-
mained at a low rate between ages seven to 12.
At age 13 there was a relative increase of activ-
ity and this delinquent activity increased more
dramatically between ages 14 to 17 with a
peaking at ages 15 and 16. Although delinquent
activity at age 17 remained relatively high, the
beginning of a falling off is observed.

Juvenile's Place of Birth-The data in Table
3.4.4 indicate that 84.6 percent of juveniles
committing first-time offense were born in the
US Virgin Islands, 5.0 percent in the US and a
combined 6.7 percent for other Caribbean ar-
eas, except Puerto Rico with 1.9 percent. In
comparison, figures for five- to 19-year-olds in
the general population of St Croix for which
data is available show persons born in St Croix
account for 74.6 percent of the population. Per-
sons born in Puerto Rico were 1.9 percent and
in the matching Caribbean areas combined 6.7
percent. The US mainland accounted for 9.9
percent (1995 Population and Housing Survey).
It must be noted that these figures represent the
entire US Virgin Islands population, and that
these groups are not necessarily spread uni-
formly across the islands.


Table 3.4.4 Juvenile's Place of Birth


Juvenile
Sex Number Percent
Total 732 100.0
United States 37 5.0
Virgin Islands 619 84.6
Puerto Rico 13 1.8
St Kitts or Nevis 7 0.9
Antigua or Barbuda 10 1.4
Dominica 8 1.1
Other Eastern Caribbean 24 3.3
Other 14 1.9


Juvenile's Race-Table 3.4.5 indicates that
98.1 percent of juveniles involved in delinquent
behavior were black and 1.1 percent were
white. Data on the age group five to 19 in the
general population (1995 Population and Hous-
ing Survey) indicated that 72.4 percent of
St Croix's population was black, 5.7 percent
white, and 21.8 percent fell in the other cate-
gory. These figures suggest that a dispropor-
tionately larger number of blacks became in-
volved in delinquent behavior, while whites
and other groups were underrepresented based
on their size in the St Croix population.

Household Status of Parents-Of the cases in
Table 3.4.6 reporting data on the presence of
parents in the household from which juvenile
delinquents reside, 86.1 percent of mothers
lived in the household compared to 27.8 per-

Table 3.4.5 Juvenile's Race


Juvenile
Sex Number Percent

Total 639 100.0
Black 627 98.1
White 7 1.1
Asian 5 0.8








PRESENTATION ANALYSIS


Table 3.4.6 Household Status of Parents


Mother Father
Household Status Number Percent Number Percent
Total 806 100.0 723 100.0
In the Household 694 86.1 201 27.8
Not in the Household 112 13.9 522 72.2

Table 3.4.7 Age of Parents

Mother Father
Age in Years Number Percent Number Percent
Total 454 100.0 213 100.0
14-19 1 0.2 0 0.0
20-29 2 0.4 0 0.0
30-39 232 51.1 55 25.8
40-49 183 40.3 102 47.9
50 & above 36 7.9 56 26.3


cent of fathers. This supports a general concept
in discourse on the causes of juvenile delin-
quency that suggest households where either
parent is absent may foster a less supportive
and stable environment that permits higher risk
to juveniles becoming delinquent. Unfortu-
nately, data are not available on the presence of
mothers and fathers in the household for the
general population of St Croix.

Age of Parents-In Table 3.4.7, fathers and
mothers of juvenile delinquents appear to be
differently represented in the various age cate-
gories. The majority of mothers-51.1 per-
cent-fell in the 30 to 39 age group, whereas
fathers were most represented in the 40 to 49
age category. Thus, fathers tended to be some-
what older than mothers.

Marital Status of Parents-Table 3.4.8 shows
data on the marital status for parents of those
juvenile delinquents. Available data show that
in the highest category, 41.4 percent of the
mothers of juvenile offenders and 57.3 percent


of the fathers of juvenile offenders were mar-
ried at the time of the juvenile's first offense,
(now married), though not necessarily to each
other. The next two highest categories for
mothers were "Never Married," 26.7, and
"Divorced," 18.0 percent. (Here again, divorce
is not necessarily between the two natural par-
ents of the juvenile). For fathers, the next two
highest categories were "Divorced," 21.9 and
"Separated," 11.1 percent.


Table 3.4.8 Marital Status of Parents

Mother Father
MaritalStatus Number Percent Number Percent

Total 589 100.0 361 100.0
Now Married 244 41.4 207 57.3
Widowed 37 6.3 3 0.8
Divorced 106 18.0 79 21.9
Separated 45 7.6 40 11.1
Never Married 157 26.7 32 8.9








PRESENTATION ANALYSIS


Juvenile's Guardian-In Table 3.4.9 of the
reported cases of juvenile delinquents that had
information on the person or persons they were
currently living with, 51.8 percent were living
with their mother, 4.2 percent with their father,
16.3 percent with mother and father, 12.5 per-
cent with their relatives, and 10.0 percent with
their mother and stepfather. These data suggest
that the majority of juvenile delinquents lived
with their mother alone. It should be noted that
the category of single female was the highest
represented category for guardian in house-
holds of the general St Croix population. To
get a more in-depth understanding of the sig-
nificance of these data, it would have been de-
sirable to have data on this same variable for
the general St Croix community.

Marital Relationship of Natural Parents-
Data on the current marital relationship of natu-
ral parents highlight categories for the natural
parents of juvenile delinquents at time of first
offense. "Never Married" and "Divorced" cate-
gories represented 40.0 and 21.5 percent re-
spectively, representing a combined total of
61.5 percent. The proportion for the category
"Married and Living Together" was 20.9 per-
cent. The fact that categories like "Both Parents

Table 3.4.9 Juvenile's Guardian

Guardian Number Percent
Total 874 100.0
Mother and Father 142 16.3
Mother 453 51.8
Father 37 4.2
Mother and Stepfather 87 10.0
Father and Stepmothel 17 2.0
Stepmother 2 0.2
Stepfather 1 0.1
Foster Parents 8 0.9
Relatives 110 12.5
Other 17 2.0


Table 3.4.10 Marital Relationship of
Natural Parents


Status Number Percent

Total 628 100.0
Married and living together 131 20.9
Separated 49 7.8
Divorced 135 21.5
Never Married 251 40.0
One parent deceased 60 9.5
Both parents deceased 2 0.3



Deceased" accounted for a low rate of juvenile
delinquents might be due to an equally small
occurrence of such cases in the general
St Croix community (Table 3.4.10).

Siblings in the Household-Data in Table
3.4.11 indicate that the lower the number of
siblings that were in the household, the greater
the representation of juvenile delinquents. For
example, in households with three or fewer sib-
lings, the range of juvenile delinquents was
16.7 to 20.5 percent of the total. In those
households with four to six or more siblings,
5.6 and 9.2 percent of the juveniles were repre-
sented. This may imply that the more siblings
present within families the more opportunities


Table 3.4.11 Siblings in the Household

Siblings Number Percent
Total 606 100.0
No siblings 101 16.7
One sibling 123 20.3
Two siblings 124 20.5
Three siblings 117 19.3
Four siblings 56 9.2
Five siblings 34 5.6
Six or more siblings 51 8.4







PRESENTATION ANALYSIS


Table 3.4.12 School Enrollment


Enrollment Status Number Percent

Total 704 100.0
Enrolled 546 77.6
Not Enrolled 158 22.4


there are for interaction with large numbers of
children and the less likely juveniles are to en-
gage in delinquent behavior.

School Enrollment-Table 3.4.12 indicates
that the majority of juvenile delinquents were
enrolled in school. Unfortunately, data indicat-
ing actual attendance at school and class par-
ticipation were not part of the data set. Al-
though 77.6 percent of juveniles involved in
delinquent behavior were enrolled in school,
that does not mean they attended school, par-
ticipated in and benefited from class activities.
It is important to note that 22.4 percent were
not enrolled in school although the law stipu-
lates that children below the age of 16 must be
enrolled in school. Attendance at school and
class is neither enforced nor easily enforceable.

Current or Highest Grade Completed-
Table 3.4.13 indicates that juveniles involved
in delinquent activities were represented across
Grades 1 through 12. However, juvenile of-
fenders were only limitedly reprepresented in
Grades one through five. There was an in-
crease in representation in Grades six to nine
with a range between 14.9 and 26.8 percent of
the juvenile offenders. The proportion of juve-
niles involved in delinquent behavior began to
decrease around Grade 10 and continued to do
so to the higher grades. It should be remem-
bered that these data represented the educa-
tional status at time of first offense and cannot
predict the educational progress after that of-
fense.


Table 3.4.13


Current or Highest
Grade Completed


Grade Number Percent
Total 586 100.0
Grade 1 2 0.3
Grade 2 1 0.2
Grade 3 4 0.7
Grade 4 2 0.3
Grade 5 9 1.5
Grade 6 87 14.9
Grade 7 157 26.8
Grade 8 106 18.1
Grade 9 102 17.4
Grade 10 53 9.0
Grade 11 42 7.2
Grade 12 21 3.6
3.4.2 St Croix: Department of Human
Services Data Relationships

Juvenile's Age by Current or Highest Grade
Completed-Table 3.4.14 presents the age of
juvenile delinquents together with the highest
grade achieved at that time. It could be as-
sumed that on average the upper age limit for
Grades one through six would be 13 years,
though it is not unusual to find children above
or below that age. Juvenile delinquents who
are in Grades one to six and between ages 14 to
17 amounted to 62.4 percent of the total. In
Grade seven, the 15- to 17-year-olds accounted
for 56.4 percent, and 51 percent in Grade eight.
Grade nine had 20.6 percent of delinquents
who were 17 years old.

Children develop at different levels, however,
but the sense of underachievement can be very
severe and lead to activity that could draw at-
tention to the low or underachiever. This trend
of a juvenile at a certain age being in a lower
grade than expected is of concern, especially
since the policy of social promotion existed
during this time in public schools.








PRESENTATION ANALYSIS


Table 3.4.14 Juvenile's Age by Current or Highest Grade


Age in Total Grade 1-6 Grade 7 Grade 8
years Number Percent Number Percent Number1 Percent Numberd Percent
Total 860 100.0 104 100.0 156 100.0 105 100
7-9 3 0.3 2 1.9 0 0.0 0 0.0
10-12 31 3.6 21 20.2 5 3.2 1 1.0
13 57 6.6 16 15.4 20 12.8 12 11.4
14 151 17.6 26 25.0 43 27.6 32 30.5
15 202 23.5 18 17.3 41 26.3 29 27.6
16 219 25.5 17 16.3 28 17.9 23 21.9
17 197 22.9 4 3.8 19 12.2 8 7.6

Table 3.4.14 Juvenile's Age by Current or Highest Grade: Continued

Age in Grade 9 Grade 10 Grade 11 Grade 12
years Number Percent Number Percent Number| Percent Numberd Percent

Total 102 100.0 102 100.0 42 100.0 21 100.0
7-9 0 0 0 0.0 0 0.0 0 0.0
10-12 1 1.0 1 1.0 0 0.0 0 0.0
13 0 0.0 0 0.0 0 0.0 0 0.0
14 17 16.7 17 16.7 1 2.4 1 4.8
15 24 23.5 24 23.5 7 16.7 0 0.0
16 39 38.2 39 38.2 10 23.8 5 23.8
17 21 20.6 21 20.6 24 57.1 15 71.4


The reverse-a juvenile in a grade above that
expected for the average age or overachiev-
ers-was largely absent. For example, there
was a lack of juvenile offenders 13 years or
younger in Grade nine or higher. Juveniles in
Grades 11 and 12, as would be expected,
tended to be 16 and 17 years old-80.9 and
95.2percent, respectively.

Juvenile's Employment Status-Table 3.4.15
shows that the majority of juveniles involved in
delinquent behavior were unemployed-84.3
percent. This could suggest a confirmation of


the old adage of mischief and idle hands. Here
again, it could help to have figures showing the
proportion of this age group of juveniles who
are employed and unemployed in the general
population in order to get a clearer picture.

Current Job-Table 3.4.16 suggests that at
time of first offense the juvenile delinquents
were employed in a variety of job types and no
particular activity seems to have had a monop-
oly on their time. It is assumed that these are
part-time jobs but it is not known if these are of
weekly or monthly duration.








PRESENTATION ANALYSIS


Table 3.4.15 Juvenile's Employment Status

Employment Status Number Percent
Total 630 100.0
Employed 99 15.7
Unemployed 531 84.3

Table 3.4.16 Current Job

Job Number Percent
Total 92 100.1
Restaurant/Hotel 4 4.4
Construction 6 6.5
Store Clerk 0 0.0
Mechanic Shop 4 4.4
Gas Station Attendant 4 4.4
Other 74 80.4



Children in the Household-Table 3.4.17
seems to suggest that households with one, two,
three or no children were more likely to have
delinquent juveniles-between 14.0 and 20.3
percent-than households with four or more
children. It is possible that the presence of more
children provide examples of alternate behav-
ior, persons with whom youngsters may associ-
ate, communicate and so limit negative outside
influences on their behavior and activities.

Juvenile's Children-Table 3.4.18 shows that
a small proportion of juvenile delinquents-2.6
percent had children of their own.

Juvenile Responsible for Childcare-Table
3.4.19 suggests that 4.6 percent of juvenile of-
fenders were responsible for childcare. This
responsibility did not necessarily refer to the
offspring of juvenile offenders, and could have
been a younger sibling or relative.

Presence of Risk Indicators-Each of the in-
dicators in Table 3.4.20 holds some signifi-


Table 3.4.17 Children in the Household

Siblings I Number Percent
Total 600 100.0
No children 84 14.0
One child 118 19.7
Two children 112 18.7
Three children 122 20.3
Four children 59 9.8
Five children 50 8.3
6 to 10 children 54 9.0
11 to 14 children 0 0.0
15 & over 1 0.2

Table 3.4.18 Juvenile's Children


Number Percent
Total 607 100.0
Children 16 2.6
No Children 591 97.4

chance for the juvenile's psychological well be-
ing. The presence of these factors contributes to
the undermining of the juvenile's self esteem,
self-worth and a healthy sense of being in con-
trol of events affecting one's life.

The influence of any one or more of these vari-
ables at any level usually suggests negative
connotations for behavior. Juvenile offenders
had relatively high percentages in the following
areas: trouble controlling anger, 62.9; frequent
arguments and fights, 58.8 suspension from

Table 3.4.19 Juvenile Responsible

for Childcare


Childcare Number Percent
Total 592 100.0

Responsible 27 4.6

Not Responsible 565 95.4







PRESENTATION ANALYSIS


Table 3.4.20 Juvenile's Psychological Assessment


[


Yes
I


No
i


I Totall Numberi Percenti Numberl Percent
Felt like a failure or worthless 448 163 36.4 285 63.6
Had trouble controlling anger 498 313 62.9 185 37.1
Had frequent arguments & fights 495 291 58.8 204 41.2
Ever been suspended from school 460 306 66.5 154 33.5
Ever been expelled from school 465 23 4.9 442 95.1
Used drugs or alcohol 415 199 48.0 216 52.0
Getting hurt (physically abused) 347 100 28.8 247 71.2


school, 66.5; use of drugs or alcohol, 48.5;
feelings of failure or worthless, 36.4; getting
hurt (physically abused), 28.8. The factor with
the lowest representation of juvenile offenders
was expulsion from school, 4.9 percent, which
may imply that delinquent activities are not
consistently curtailed in school or that the rules
and regulations are not that severe.

Juvenile's Sexual Activity-Table 3.4.21 indi-
cates that more than half-57.9 percent-of the
first-time offenders were sexually active. This
variable is often identified as an indicator of
association with delinquent behavior.

Forced Sex-Table 3.4.22 indicates that 9.4
percent of the juveniles with first offenses were
forced to have sex. When these data are exam-
ined along with the previous table, Juvenile's
Sexual Activity, the data suggest that a large
proportion of juveniles opt to participate in sex-
ual activity freely. It must be remembered,
however, that a juvenile's perception of free
choice may not meet all psychological stan-
dards and youngsters may be under subtle pres-
sures and constraints that propel them into
negative sexual activity.

Juvenile's Sex Partners-In Table 3.4.23 the
number and percentage of juveniles for whom
data are available here is relatively low, given
that 124 offenders had indicated they were


Table 3.4.21 Juvenile's Sexual Activity

Sexual Activity Number Percent
Total 214 100.0
Sexually Active 124 57.9
Not Sexually Active 90 42.1


Table 3.4.22 Forced Sex


INumber Percent
Total 223 100.0
Forced to have sex 21 9.4
Not forced to have sex 202 90.6


sexually active. These data
reliable.


may not be very


Pre-outcome Status-Of the six offenses
listed in Table 3.4.24, the number of juveniles
involved increasingly decreased in number, ex-
cept for Offense 6 with an increase above the
offenses immediately preceding. The propor-
tion of cases going to trial rather than plea
agreement was relatively low for all offenses.
As the cases spread over the six offenses, there
does not seem to be any definite pattern indi-
cating an increase or decrease in the proportion





I







PRESENTATION ANALYSIS


Table 3.4.23 Juvenile's Sex Partners


Sexual Partners Number Percent
Total 17 100.0
1 15 88.2
2 1 5.9
3 1 5.9
4 0 0.0
5 0 0.0
6 to 10 0 0.0
10 and over 0 0.0


of cases that went to trial as against the propor-
tion that utilized a plea agreement.

Type of Offense-Table 3.4.25 lists the type of
offenses committed by juvenile delinquents for
a total of six offenses. Felonious Assaults,
Other Part I Felony and Other Part II Felony
account for the highest number of cases among
offenses.
Overall, the type of offense or crime varied
from one offense to another. Whereas offenses
committed by juvenile delinquents in the early
offenses seemed to be spread widely, offenders
with four or more offenses seemed to have nar-
rowed down their scope to activities such as


robbery, felonious assault, burglary, possession
of a firearm and Other Part II Felony. It is pos-
sible that these delinquents are not just experi-
menting but have settled into a pattern of more
serious activities.

Year of Offense-Table 3.4.26 indicates the
year that juvenile delinquents committed their
first through sixth offenses. Although the time
period of the study examined cases current be-
tween 1987 and 1997, a number of cases had
first offenses occurring before the study's time
frame and the cases remained active during the
period covered by the study.

Cases were on record from as early as 1980. It
should be noted that since juvenile delinquent
behavior is recorded for 10 years between ages
7 to 17, the year 1997 did not necessarily indi-
cate the end of the case history of all cases cov-
ered in the study. Only those delinquents with a
first offense in 1987 or before could safely be
regarded as having completed their history of
delinquent behavior.

Outcome of Offense-Table 3.4.27 gives a
breakdown of the outcome of cases that went
through the YIB and DHS systems for the six
offenses being tracked. This was the final proc-


Table 3.4.24 Pre-Outcome Status


Offense 1 Offense 2 Offense 3
Status Number Percent Number Percent Number Percent
Total 507 100.0 110 100 51 100
Trial 33 6.5 6 5.5 5 9.8
Plea Agreement 474 93.5 104 94.5 46 90.2

Table 3.4.24 Pre-Outcome Status: Continued

Offense 4 Offense 5 Offense 6
Status Number Percent Number Percent Number Percent
Total 6 100 6 100 21 100
Trial 1 16.7 0 0.0 3 14.3
Plea Agreement 5 83.3 6 100.0 18 85.7








PRESENTATION ANALYSIS


Table 3.4.25 Type of Offense

Offense 1 Offense 2 Offense 3
Offense type Number Percent Number Percent Number Percent

Total 888 100.0 196 100.0 87 100.0

Robbery 71 8.0 16 8.2 10 11.5
Felonious Assaults 159 17.9 30 15.3 11 12.6
Burglary 101 11.4 17 8.7 7 8.0
Grand Larceny 79 8.9 19 9.7 7 8.0
Other Part I felony 126 14.2 32 16.3 13 14.9

Possession of controlled substance 44 5.0 12 6.1 9 10.3
Possession of unlicensed firearm 37 4.2 11 5.6 7 8.0
Possession of stolen property 19 2.1 3 1.5 2 2.3
Unauthorized use of a vehicle 2 0.2 2 1.0 0 0.0
Other Part II felony 45 5.1 24 12.2 9 10.3

Aggravated assault and battery 31 3.5 5 2.6 2 2.3
Petit Larceny 12 1.4 2 1.0 0 0.0
Run away minor 1 0.1 0 0.0 0 0.0
Missing minor 3 0.3 0 0.0 0 0.0
Other 158 17.8 23 11.7 10 11.5

Table 3.4.25 Type of Offense: Continued

Offense 4 Offense 5 Offense 6
Offense type Number Percent Number Percent Number Percent
Total 8 100.0 12 100.0 40 100.0
Robbery 1 12.5 1 8.3 9 22.5
Felonious Assaults 0 0.0 3 25.0 7 17.5
Burglary 1 12.5 0 0.0 2 5.0
Grand Larceny 1 12.5 0 0.0 2 5.0
Other Part I felony 0 0.0 1 8.3 4 10.0
Possession of controlled substance 0 0.0 1 8.3 5 12.5
Possession of unlicensed firearm 1 12.5 0 0.0 2 5.0
Possession of stolen property 0 0.0 0 0.0 0 0.0
Unauthorized use of a vehicle 0 0.0 0 0.0 0 0.0
Other Part II felony 3 37.5 1 8.3 6 15.0
Aggravated assault and battery 0 0.0 0 0.0 1 2.5
Petit Larceny 0 0.0 0 0.0 0 0.0
Run away minor 0 0.0 0 0.0 0 0.0
Missing minor 0 0.0 0 0.0 0 0.0
Other 1 12.5 5 41.7 2 5.0









PRESENTATION ANALYSIS


Table 3.4.26 Year of Offense


Offense Year
Total
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998


Nu


I Offense 1
Number Percent
882 100.0
1 0.1
0 0.0
1 0.1
1 0.1
1 0.1
1 0.1
3 0.3
79 9.0
94 10.7
59 6.7
78 8.8
66 7.5
81 9.2
51 5.8
91 10.3
104 11.8
102 11.6
67 7.6
2 0.2


Table 3.4.26 Year of Offense: Continued


Offense 4
Number Percent


Offense Year


100.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
12.5
0.0
12.5
12.5
12.5
12.5
12.5
0.0
0.0
12.5
12 5


Offense 5
NumberI PercentI
11 100
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
1 9.1
0 0.0
1 9.1
2 18.2
1 9.1
1 9.1
0 0.0
2 18.2
2 18.2
1 9.1
0 0.0


Offense 6
NumberI Percent
35 100.0
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
3 8.6
4 11.4
5 14.3
3 8.6
3 8.6
1 2.9
4 11.4
5 14.3
3 8.6
3 8.6
1 2 9


Total
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998


Nu


Offense 2 Offense 3
mber| Percent Number Percent
193 100.0 85 100.0
0 0.0 0 0.0
0 0.0 0 0.0
0 0.0 0 0.0
0 0.0 0 0.0
0 0.0 0 0.0
0 0.0 0 0.0
1 0.5 1 1.2
7 3.6 1 1.2
22 11.4 11 12.9
23 11.9 6 7.1
11 5.7 7 8.2
28 14.5 8 9.4
20 10.4 10 11.8
9 4.7 5 5.9
15 7.8 7 8.2
24 12.4 11 12.9
14 7.3 13 15.3
12 6.2 2 2.4
7 3.6 3 3.5


I


-~~-~~-- ~--~~









PRESENTATION ANALYSIS


Table 3.4.27 Outcome of Offense


I Offense 1
Outcome Number Percent
Total 888 100.0
Crisis Stabilization Center 5 0.6
Youth Rehabilitiation Center 43 4.8
Boys/Girls Group Home 3 0.3
Transfer to Correctional Facility 4 0.5
Drug Rehabilitation Center 2 0.2
Psychological Evaluation 11 1.2
Dismissed 187 21.1
Left Jurisdiction 0 0.0
Case not prosecuted 70 7.9
Case pending 0 0.0
Community Service 3 0.3
Counseling 2 0.2
Restitution = $2000 10 1.1
Attend School Regularly 0 0.0
Academic Program 0 0.0
Read/Submit book report 0 0.0
Curfew/prnt/princ 2 0.2
reprimand/prin/j off 0 0.0
probation 328 36.9
Other 218 24.5
Table 3.4.27 Outcome of Offense: Continued

Offense 4
Outcome Number Percent
Total 9 100.0
Crisis Stabilization Center 0 0
Youth Rehabilitiation Center 2 22.2
Boys/Girls Group Home 0 0.0
Transfer to Correctional Facility 0 0.0
Drug Rehabilitation Center 0 0.0
Psychological Evaluation 0 0.0
Dismissed 1 11.1
Left Jurisdiction 0 0.0
Case not prosecuted 1 11.1
Case pending 0 0.0
Community Service 0 0.0
Counseling 0 0.0
Restitution = $2000 0 0.0
Attend School Regularly 0 0.0
Academic Program 0 0.0
Read/Submit book report 0 0.0
Curfew/prnt/princ 0 0.0
reprimand/prin/j off 0 0.0
probation 3 33.3
Other 2 22.2


Offense 2 Offense 3
Number Percent Number Percent
196 100.0 85 100.0
0 0.0 0 0.0
37 18.9 20 23.5
0 0.0 1 1.2
2 1.0 1 1.2
1 0.5 0 0.0
3 1.5 1 1.2
47 24.0 19 22.4
0 0.0 0 0.0
7 3.6 0 0.0
0 0.0 2 2.4
1 0.5 0 0.0
0 0.0 1 1.2
0 0.0 0 0.0
0 0.0 0 0.0
0 0.0 0 0.0
0 0.0 0 0.0
0 0.0 0 0.0
0 0.0 0 0.0
69 35.2 27 31.8
29 14.8 13 15.3


Offense 5
Number Percent
12 100.0
0 0.0
5 41.7
0 0.0
0 0.0
1 8.3
0 0.0
2 16.7
0 0.0
0 0.0
1 8.3
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
3 25.0
0 0.0


Offense 6
Number Percent
38 100.0
0 0.0
13 34.2
0 0.0
1 2.6
0 0.0
0 0.0
12 31.6
0 0.0
0 0.0
2 5.3
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
0 0.0
5 13.2
5 13.2







PRESENTATION ANALYSIS


essing of juvenile delinquents and represented
the final sentencing stage. Between offenses
one and four the most common form of treat-
ment is probation ranging between 32.2 to 36.9
percent of cases. For cases with five and six
offenses, however, the treatment in the majority
of instances changed to sentencing for a period
at the Youth Rehabilitation Center (YRC)-
between 22.2 to 41.7 percent of cases.

It should be noted that for this population of
juvenile delinquents for whom the literature
indicates there is high marijuana usage, drug
rehabilitation is listed as a treatment only in
Offenses 1, 2 and 5 and that at an extremely
low rate.

Biological Father Known-Table 3.4.28 was
inserted into the St Croix data gathering form
to measure the extent to which juvenile delin-
quents did not know the identity of their fa-
thers. In discussions among DHS social work-
ers, it was their perception that this factor
would rate high. However, these data indicate
that an overwhelming majority of juvenile de-
linquents did have knowledge of their fathers.
Perhaps closer examination should be given to
the level and types of bonds and relationships
existing between father and juvenile delin-
quents, as a more important determining factor
of delinquent behavior.



Table 3.4.28 Biological Father Known

Knowledge of Father Number Percent

Total 652 100.0

Known 609 93.4

Unknown 43 6.6


3.4.3 Predicting Recidivism with the
St Croix Department of Human
Services Data

In this analysis, five predictor variables that
showed statistically significant association with
the outcome variable, recidivism were included
in the model. The estimation of the predictor
variables was severely hampered by the wide-
spread occurrences of missing data values.
While there were 888 cases in the original data
set of St Croix's Department of Human Ser-
vices' data, only 361 valid observations could
be used in deriving the estimates. This indi-
cates that only 361 cases had non-missing val-
ues for each of the five variables that were in-
cluded in the model. The five predictor vari-
ables were Age, Sex, Status of Natural Parents,
Feeling of Failure, and Use of Drugs.

3.4.3.1 Estimates of the St Croix Department
of Human Services Data

The likelihood ratio chi-square in Table 3.4.3.1
suggests that at least one of the variables in the
model is significant. The table indicates that
while four of the variables were statistically
significant-Sex, Feelings of Failure and Use
of Drugs-two of them failed to reach the tra-
ditional level of statistical significance of 0.05.
However, because the variable Age has shown
itself to be such a reliably strong predictor in all
other models included in this study, it was re-
tained even though its significance level of
0.0612 is just slightly beyond the usual 0.05
level. When the variables Age-Squared and
Age were included in the model, both showed
no statistical significance.4 Even more surpris-
ing is the failure of the Status of Natural Par-
ents variable to reach the significance level of
0.05. Even though a 2 x 2 table cross-classified
this variable with recidivism and showed them


4A second possible explanation of the relatively poor predictive
strength of the Age variable is that the data reflect some degree of
contamination. The data analysts were unable to verify the data
as they had no direct access to the juveniles' case files.







PRESENTATION ANALYSIS


Table 3.4.3.1 Parameter Estimates in Logistic Regression Predicting Recidivism:
St Croix, Department of Human Services Data

Parameter Standard Pr> Adjusted
Variable Estimate Error Chi Square Odds Ratio
(1) (2) (3) (4)
Intercept 1.3319 1.2519 0.2874

Age -0.1545 0.0825 0.0612 0.857
Sex (1=Male) 0.4543 0.2086 0.0295 2.481

Parents (1=Married and Living -0.2322 0.1736 0.1811 0.628
Together)
Feelings of Failure (1=Yes) 0.2935 0.1267 0.0205 1.799

Drug Use (1=Yes) 0.5238 0.1223 <0.0001 2.851
Likelihood Ratio Chi Square = 48.8380 df = 5 Pr > Chi Square < 0.0001

Notes: N=361. Odds ratios greater than 1.0 indicate greater odds of recidivism.


to be significantly related, this variable was
barely below the 0.2 level of significance-
0.1811. Because of the strong association
shown in the 2 x 2 case, this variable was also
retained. It would appear that the high attrition
of cases mentioned above is probably the pri-
mary reason for the relatively poor showing of
these two variables in terms of statistical sig-
nificance.

The adjusted odds ratio for age of 0.857 may be
interpreted as follows. For each one-year in-
crease in age, the odds of a repeat offense are
decreased by a factor of 0.857 or eventuates in
a 100 x (0.857 1) = -14.3 percent change, i.e.,
14.3 percent reduction in the odds of the juve-
nile becoming a repeat offender. The odds ra-
tio for Sex of 2.481 again indicates that the
odds of males being repeat offenders are about
2.5 times higher than the odds for females.
First-time offenders who live with their natural
parents in a married state and living together
have odds that are reduced by 37.2 percent of
those juveniles who live with a parent or par-
ents in some other kind of relationship. This


may be explained in other words by saying that
first-time offenders who do not live with their
natural parents in a married union have odds of
being recidivists by a factor that is-1/0.628 =
1.591-about 1.6 times higher than the odds for
offenders who live with their natural parents in
a married relationship.

Juvenile offenders who indicated they harbored
feelings of being a failure or of being worthless
had predicted odds of being repeat offenders of
1.799. In other words, the odds of recidivism
for a first-time offender who experienced these
feelings of failure and worthlessness are about
80 percent higher than the odds for other juve-
nile offenders who do not have such feelings.

The final variable is one that derives from re-
sponses to the question concerning the use of
drugs or alcohol before or during school. The
adjusted odds ratio shown in table 3.4.3.1 of
2.851 is the highest among all the variables.
Hence, the odds of recidivism for juveniles
who indicated they used alcohol or drugs be-
fore or during school was nearly three times as








PRESENTATION ANALYSIS


Table 3.4.3.2 Predicted Probabilities in Logistic Regression: St Croix, Department of Human Services Data

Proba-
Coefficients: a Agefo Sex Parents Failure Drugs Odds ability
(1) (2) (3) (4) (5) (6) (7) (8)
Parameter
Estimates: 1.3319 -0.1545 0.4543 -0.2322 0.2935 0.5238

Juvenile Profiles:


D1 8 1 0 1 1


D2

D3

D4

D5

D6

D7

D8

D9

D10

D11

D12

D13

D14

D15

D16

D17

D18

D19

D20

D21

D22


3.926

3.364

2.882

2.469

2.116

1.813

1.553

1.331

1.141

0.977

1.055

1.082

1.343

1.065

1.253

0.934

0.625

0.993

0.787

0.788

0.630

1.707


Variables:
Agefo = Age of juvenile at first offense.
Sex = Sex of the juvenile (1 = male, 0 = female).
Parents = Natural parents of juvenile (1 = married and living together, otherwise 0).
Failure = Feeling like a failure or worthless (1 = failure, otherwise 0).
Drugs = Use of drugs or alcohol before or during school (1 = use, otherwise 0).


0.797

0.771

0.742

0.712

0.679

0.645

0.608

0.571

0.533

0.494

0.513

0.520

0.573

0.516

0.556

0.483

0.385

0.498

0.440

0.441

0.387

0.631







PRESENTATION ANALYSIS


high as the odds for those juveniles who indi-
cated they did not. Or it might be said that this
latter group has odds of being recidivists of
only about 35 percent of the odds of those us-
ing drugs or alcohol in school.

3.4.3.2 The Prediction of Recidivism

Table 3.4.3.2 shows the predicted chances of
becoming a recidivist for a combination of
stated levels of the variables in the model. The
first 10 profiles are those of juveniles for each
successive age from 8 to 17 with the other vari-
ables in the model-Sex, Status of Natural Par-
ents, Feeling of Failure, and Use of Drugs-
remaining constant throughout the age range
(See Figure 3.4.3.1). It is noted that for these
predictions shown in column (8), there is a
gradual decrease from a high of 0.797 to a low
of 0.494. The first profile may be interpreted to
say that an 8-year-old first-time male offender


who is living with parents that are not married
and living together, who thinks of himself as a
failure or worthless and who used drugs or al-
cohol before or during school is predicted to be
an offender who will be arrested at least one
more time before he reaches the age of 18. The
data show that this combination of characteris-
tics predicts recidivism for all ages up to 16
years. It is not until the juvenile reaches 17
years in age that his prediction falls below 0.50,
therefore indicating that he will likely not com-
mit another offense before he is 18 years old.
Profile Dll suggests that a 15-year-old-male
living with his natural parents who are married
but who has low self esteem, and has used
drugs or alcohol at school is still predicted to be
a repeat offender before age 18. Profiles Dl to
D22 in the table present other combinations of
characteristics and their predictions of repeat or
non-repeat of offenses that result in arrests.


Figure 3.4.3.1 Predicted Chances of Recidivism: St Croix DHS
For a Male Using Drugs, With Feelings of Failure and not Living With Married Parents


0.75 -


0.60--
C-)
0 0.45--
0





0.15


0.00
0 .0 0 -- - -
D1 D2 D3 D4 D5 D6 D7 D8 D9 D10
Note: See text for profile characteristics Juvenile Profile




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs