Close
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Psychopathy, psychosis, drug abuse, and reoffense among conditionally released offenders
(USC Thesis Other)
Psychopathy, psychosis, drug abuse, and reoffense among conditionally released offenders
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
PSYCHOPATHY, PSYCHOSIS, DRUG ABUSE,
AND REOFFENSE AMONG CONDITIONALLY
RELEASED OFFENDERS
by
Frederick M. Blum
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
EDUCATION (COUNSELING PSYCHOLOGY)
August 2003
Copyright 2003 Frederick M. Blum
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3116668
Copyright 2003 by
Blum, Frederick M.
All rights reserved.
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 3116668
Copyright 2004 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90089-1695
This dissertation, written by
‘ L D < ‘ 4^ M QUIA A A -_ _ _ _ _ _ _ _ _
under the direction o f h is dissertation committee, and
approved by all its members, has been presented to and
accepted by the Director o f Graduate and Professional
Programs, in partial fulfillment o f the requirements fo r the
degree o f
D O C TO R O F PHILO SO PHY
Director
Date A u gu st 1 2 , 2003
Dissertation Committee
Chair
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DEDICATION
With heartfelt gratitude I dedicate this research to my wife, Cynthia;
our son, Thomas; and Grandma Betty Voight. They have been co-laborers
and have endured my endeavors through the thick and thin of all that has
been a part of this project. To Cynthia, who has never stopped believing
and giving support to me, both emotionally and in actual participation in the
data gathering as a research colleague: I hold you in high honor. To
Thomas, my son, who was conceived and raised on this project and who
himself greatly delights in the culmination of this effort: I salute you and
praise you for your patience and endurance to “ wait” for your “ tree house” to
be completed after this research is finished. To Grandma, who “picked up
the pieces” at each turn in the journey and was always a positive encour
aging influence: Thanks so much.
I also wish to make a special mention of appreciation and dedication
to both of my parents, Charles and Mary Blum, who believed in my vision of
completion and were strong support figures during the early stages of my
study. They both passed away during the last year of this journey. Mom
and Dad, you are missed but not forgotten. Thank you for all that you be
lieved in and gave to me to make this work possible to complete.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACKNOWLEDGMENTS
I am deeply grateful for the encouragement, patience, and sugges
tions given to me by my mentors, bosses, and colleagues at the CONREP
clinic. Special thanks go to my bosses and friends, Louis Hodnett, M.S.W.,
CONREP Director for San Bernardino/Riverside County; and Munir Sewani,
Ph.D., CONREP Clinic Supervisor.
Robert Mone, State Database Manager, was always available and it
was a delight to work with him. Thanks so much, Bob, for your unending
data searches and frequent delving into the database to fill my requests.
Gary Gleason, Director of HIMD at Patton State Hospital, and his
appointed staff, Patricia Williams, HIMD Files Supervisor, and Ingrid
Roberts, HIMD Files OA, were especially helpful in locating files and mak
ing the physical task of transporting files to and from the examining rooms a
smooth and hassle-free task. I thank them so much for this assistance.
William Summers, Executive Director, Patton State Hospital, and
Bruce Parks, Executive Director, Patton State Hospital, gave their full sup
port and were helpful in facilitating the actual beginning of the data collec
tion process. Brad Smith, Chief, CFL Office, Patton State Hospital, gave
ready answers and solutions to the unforeseen problems of many kinds as
they arose.
Jane Goerss, Ph.D., Chair, Research Committee, Patton State Hos
pital, was helpful and encouraging as she presented suggestions and advo
cated for this research to begin. I thank her.
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Thanks to my research assistants for their persistence and enthusi
asm for research and collegial camaraderie: Cynthia Blum, M.S., M.F.T.,
my loving and faithful companion; Jilayne Gouvion, Ph.D., colleague and
friend; Meredith Smith, Ph.D., colleague and friend; Elizabeth Ramos, M.A.,
psychology intern; David Chu, M.A., psychology intern; and Alisa Lite, M.A.,
psychology intern.
I greatly appreciate the insightful and helpful editing comments and
guidance of each of my committee members: John Brekke, Ph.D.; Rodney
Goodyear, Ph.D.; and Michael Newcomb, Ph. D., my committee chair.
Special appreciation goes to my current and former CONREP col
leagues. I also wish to thank my extended family and friends who have
supported and encouraged my efforts with this research.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF CONTENTS
Page
DEDICATION................................................................................................. ii
ACKNOWLEDGMENTS............................................................................... iii
LIST OF TABLES AND FIGURES............................................................... viii
ABSTRACT.................................................................................................... ix
Chapter
1. INTRODUCTION............................................................................. 1
Statement of the Problem ....................................................... 2
Purpose and Objectives.......................................................... 4
Rationale.................................................................................... 4
Research Hypotheses.............................................................. 5
2. LITERATURE REVIEW .................................................................. 6
Recidivism ................................................................................ 6
Problems with Definitions and Data Collection .................... 7
General Male Prison/Jail Population...................................... 9
The 1983 BJS Report on Recidivism.............................. 9
Comparison of Recidivism Rates for Prisoners
Released in 1983 and 1994 ...................................... 11
Meta-Analytic Studies of Recidivism .............................. 13
Spouse Abuse Offenders........................................................ 16
Sex Offender Studies............................................................... 17
Substance Abuse and Dual Diagnosis Offenders................ 20
Prison Sam ples................................................................. 20
Community Studies .......................................................... 22
Adolescent Offenders....................................................... 26
Foreign Offender Populations.......................................... 27
Finland Offenders ...................................................... 27
Canadian Offenders................................................... 27
Nordic Offenders........................................................ 28
Civilly Committed and Insane Offenders............................... 28
Summary of General Male Recidivism Studies.................... 32
The Three Hypothesized Predictors and Offender
Studies................................................................................ 32
Psychopathy Studies Using the Psychopathy
Checklist, Screening Version (PCL-R:SV) ............. 32
General Male Offender Studies of Psychopathy 33
Sex Offenders and Psychopathy...................................... 35
Schizophrenic Offenders and Psychopathy.................... 35
Active Psychosis Studies.................................... 36
Substance Abuse and Mental Illness Studies...................... 39
v
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter Page
Summary of Hypothesized Predictor Literature ................... 40
Theoretical Limitations............................................................ 41
Relevant Theories for This Study ................................... 41
Dopamine Hypothesis ...................................................... 43
Correctional Treatments That Work ............................... 44
The Biopsychosocial Model and Violence............................. 45
Biopsychosocial Model and Active Psychosis .............. 45
Biopsychosocial Model and Substance Abuse
Problems..................................................................... 47
Biopsychosocial Model and the Presence of
Psychopathy ............................................................... 48
The Psychopathy Construct.................................................... 51
Psychopathy History......................................................... 53
Epidemiology of Psychopathy ......................................... 57
Schizophrenia/Psychosis and Violence......................... 58
Dual Diagnosis and Violence........................................... 61
Summary of Hypothesized Predictor Literature ................... 63
3. M ETHOD.......................................................................................... 65
Participants............................................................................... 65
Population Parameters..................................................... 67
Legal Status................................................................ 67
Diagnosis.................................................................... 67
Age .............................................................................. 68
Exposure to Risk........................................................ 69
Gender Issues............................................................ 69
File Selection..................................................................... 69
Data Collection and Measurements....................................... 70
Instrumentation.................................................................. 70
PCL:SV........................................................................ 71
H C R -20....................................................................... 74
Raters and Rater Training ............................................... 76
Coding Criteria for Independent Variables..................... 78
Coding Outcome Dependent Variables......................... 80
Definition of Term s................................................................... 81
4. RESULTS......................................................................................... 83
Descriptive Statistics of the Sample....................................... 83
Preliminary Group Comparisons............................................ 88
Hypothesis 1 ...................................................................... 91
Hypotheses 2 and 3 ......................................................... 92
Hypothesis 4 ...................................................................... 92
5. DISCUSSION.................................................................................. 98
Summary of the Study ............................................................ 98
Discussion of the Hypotheses................................................ 99
Hypothesis 1: Psychopathy............................................ 99
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter Page
Hypothesis 2: Active Psychosis..................................... 102
Hypothesis 3: Substance Abuse.................................... 105
Hypothesis 4: Comparison of Predictive Power of
the Instruments .......................................................... 109
Implications for Explanatory Theories................................... 113
Implications for Clinical Practice ............................................ 117
Limitations of the Study........................................................... 119
Recommendations for Future Research ............................... 121
REFERENCES.............................................................................................. 123
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF TABLES AND FIGURES
Table Page
1. Independent and Dependent Variables Used in This Study ............ 3
2. Coding Criteria for the Independent Variables.................................. 79
3. Demographic and Diagnostic Characteristics of Sample
(A/= 53) .............................................................................................. 84
4. Means and Standard Deviations of the Predictor Variables
for Three Outcome Groups............................................................... 89
5. Results of t Tests of the Mean Differences Between Violent and
Nonviolent Re-Offense Groups for the Predictor Variables
Psychopathy Psychopathy Checklist: Screening Version
Total Score, Historical, Clinical-15 Total Score, and the Three
Other Predictors Psychopathy, Active Psychosis, and
Substance Abuse .............................................................................. 90
6. Results of t Tests of the Mean Differences Between Successes
and Violent/ Nonviolent Re-Offense Groups for the Predictor
Variables Psychopathy Psychopathy Checklist: Screening
Version Total Score, Historical, Clinical-15 Total Score, and
the Three Other Predictors Psychopathy, Active Psychosis,
and Substance Abuse........................................................................ 91
7. Coordinates of the Curve...................................................................... 96
8. Means and Standard Deviations for Violent and Nonviolent
Groups (N = 2 5 )................................................................................. 111
9. Means and Standard Deviations for Success Group (N = 13) ......... 112
Figure
1. The test result variable(s): Psychopathy Checklist: Screening
Version (PCL:SV) Total Score with Area Under the Curve
(AUC) = .678; and Historical, Clinical-15 (HC-15) Total Score
with AUC = .727.................................................................................. 95
viii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
Records of 53 male schizophrenic state hospital patients, ages 23 to
38, were coded for psychopathy, active psychosis, and substance abuse as
predictors of Conditional Release Outcome in the State of California’s Con
ditional Release Program (CONREP) during a follow-up period of up to 7
years. All but one patient had a violent criminal history. The Psychopathy
Checklist: Screening Version (PCL:SV) and the Historical and Clinical and
Risk Management-20 (HCR-20, shortened to HC-15) were used to code the
predictors.
The primary purpose of the study was to determine the predictive
correlation of the presence of psychopathy, active signs of psychosis, and
substance abuse problems with violent and nonviolent reoffense among
schizophrenic patients released from a state hospital as Guilty, but Not
Guilty by Reason of Insanity (NGRI) of serious felony offenses.
The secondary purpose of the study was to determine whether dy
namic or changeable clinical factors, as assessed on the Clinical or “C”
scale of the HCR-20, improved the risk assessment of static or unchanging
criminal behavior history and durable psychopathic personality factors as
assessed by the PCL.SV.
The hypothesis that psychopathy predicts future violent and nonvio
lent offenses was supported at the p = .01 level of significance. The hy
potheses that active signs of psychosis and substance abuse problems
would predict reoffense were not supported. The overall comparison be
tween the PCL:SV and the HC-15 showed that both assessment tools
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
predicted reoffense. The HC-15 predicted violent and nonviolent reoffense
at a significance level of p = .01. The PCL:SV predicted violent and nonvio
lent reoffense at a significance level of p = .06. The area under the curve
(AUC) for the PCLSV was .678 for violent and nonviolent reoffenses. The
AUC for the HC-15 was .727, predicting more of the violent and nonviolent
reoffenses than the PCLSV.
In this study, 40 of 53 (75.5%) patients reoffended, and 15 (28.3%)
of those reoffended violently. Suggestions for a biopsychosocial assess
ment model to include biological, psychological, and social factors in as
sessment and treatment of schizophrenic offenders are discussed, as well
as suggestions for future research.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 1
INTRODUCTION
Mental health professionals make decisions daily that affect public
safety. Among these is the assessment of the degreed risk of violent
reoffense for living in the community. In California, mental health staff in
the Conditional Release Program (CONREP) supervise individuals who
have been found “Guilty, but Not Guilty by Reason of Insanity” (NGRI) of
felonious and violent crimes. Entry into CONREP is established only after
clients are court ordered to be released from state hospitals and to enter
outpatient treatment. The court orders these offenders to make the transi
tion to Community Outpatient Treatment (COT) based on the risk assess
ments provided by hospital and community program clinicians.
Research indicates that risk assessments made using primarily clini
cal judgment have shown very low accuracy in predicting dangerousness in
forensic hospital patients (Steadman & Cocozza, 1979; Thornberry &
Jacoby, 1979). More recently, a few studies have used assessment tools
that improved prediction accuracy regarding individuals who are at risk for
violence and reoffense (Douglas, 1996; Douglas, Ogloff, Grant, & Nicholls,
1999; Harris, Rice, & Quinsey, 1993; Hart, Hare, & Forth, 1994). In the
past, “ expert” clinicians relied on currently observed behavior and often
ignored historical data. These newer devices utilize a broad range of
behavior data, including criminal histories, personality characteristics, and
affective qualities.
1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Paraprofessional and professional clinical staff need more accurate
predictive assessment tools and procedures when making outpatient
readiness recommendations to the court. The Historical, Clinical and Risk
Management-20 (HCR-20), one of the measures in increasing use among
forensic evaluators, also includes the interpretation of current and observ
able dynamic mental status, attitudes, and behaviors.
Two devices were utilized in this study: (a) the Psychopathy
Checklist: Screening Version (PCLSV), and (b) the HCR-20. (More detail
about these instruments is provided later in this study.) Predictive assess
ment tools are urgently needed to assist clinicians in making appropriate
outpatient readiness recommendations to the court. However, no published
studies report the use of these specific instruments with the CONREP
population.
In this retrospective study, CONREP patients’ state hospital records,
CONREP records, arrest records, and rehospitalization records were
examined to code the hypothesized predictive markers for the risk of
general and violent reoffense in individuals conditionally released by the
court into the CONREP program of the California State Department of
Mental Health (DMH).
Statement of the Problem
This study addresses the hypothesized link between general and
violent reoffense (see glossary for definitions) and the presence or absence
of three independent variables: (a) psychopathy (as assessed by the
PCLSV), (b) active psychosis (delusions and hallucinations), and (c) history
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
of drug/alcohol abuse. The study’s central focus is the prediction of the risk
for general and violent reoffense (dependent variables) of California’s
Conditionally Released Offenders supervised by the CONREP Program.
These offenders are from one legal class— PC1026 (NGRI); they are also
from the diagnostic groups Schizophrenia and Schizoaffective Disorder.
The variables are summarized in Table 1.
Table 1
Independent and Dependent Variables Used in This Study
Independent variables Dependent variables
(predictor variables) (outcome variables)
1. Psychopathy, coded from
item 7 on the Historical
Scale of the Historical,
Clinical-15
2. Signs of active psychosis,
coded on item 3 of the
Clinical Scale of the
Historical, Clinical-15
3. Substance abuse problems,
coded on item 5 of the
Historical, Clinical-15
4. Pyschopathology Checklist:
Screening Version total score
5. Historical, Clinical-15
total score
Note. Historical, Clinical-15 = the Historical, Clinical, and Risk Manage
ment-20 instrument, excluding the five Risk Management items; CONREP =
Conditional Release Program of the California Department of Health.
3
1. Success (n = 13): those who
remain on CONREP or “ free” in the
community
2. Violent and nonviolent reoffenders
(n = 40)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Purpose and Objectives
The purpose of the study is to examine the relationships among psy
chopathy (as defined by the PCLSV), active psychosis, history of alcohol
and drug abuse, and general or violent reoffense in the community. The
specific objectives of the study are: (a) to demonstrate that psychopathy
versus nonpsychopathy is a predictive and measurable construct with
regard to mentally ill offenders with psychosis (with schizophrenia as the
modal diagnosis); (b) to show the contribution that the other independent
variables (i.e., drug/alcohol abuse and active psychosis) have on the pre
diction of general and violent reoffense in this population; (c) to show that
static/historical and dynamic/current mental status data combine to make
more accurate predictions of general and violent reoffense in the studied
population of psychotically diagnosed offenders; and (d) to demonstrate
with these instruments a higher degree of accuracy in predicting general
and violent reoffense.
Rationale
The rationale for this study comes from the need of mental health
professionals to assess more accurately the outpatient readiness status of
forensic patients and to communicate these assessments as recommenda
tions to the courts. Past research has shown that these types of clinical
expert opinions frequently show very low accuracy in predicting dangerous
ness and reoffense (Steadman & Cocozza, 1974; Thornberry & Jacoby,
1979). More recently, there have been studies with newly developed
measuring tools showing improved prediction accuracy regarding violence
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and reoffense (Douglas, 1996; Douglas et al., 1999; Hart et al., 1994). In
the past, clinicians relied on present, observable behavior and often ignored
historical data. These newest devices focus heavily on historical data,
including criminal histories as well as personality and affective traits. The
HCR-20 also includes the interpretation of current and observable dynamic
mental status, attitudes, and behaviors.
Research Hypotheses
In reference to the assumptions and discussion above, four hypothe
ses were tested.
1. Violent and nonviolent reoffenders will have higher psychopathy
then nonreoffenders.
2. Violent and nonviolent reoffenders will show signs of active
psychosis more than nonreoffenders.
3. Violent and nonviolent reoffenders will have more substance
abuse problems than nonreoffenders.
4. The 15 items on the Historical and Clinical scales of the HCR-20
will more accurately predict violent and nonviolent reoffense in this
population than will the PCLSV.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 2
LITERATURE REVIEW
The literature review addresses recidivism of violent and general (or
nonviolent) offense in several populations, followed by a review of the three
variables hypothesized to predict criminal reoffense among conditionally
released schizophrenic patients: psychopathy, active psychosis, and drug
abuse history. Next, the biopsychosocial model (Engel, 1977), which
integrates three theory “ fields" for assessment and treatment, is presented,
followed by some “mini” theories explaining the hypothesized predictor
variables.
The recidivism review starts by identifying some limits and problems
encountered with the definition and use of the word recidivism in the litera
ture. The definition of recidivism changes with the context of the particular
study of reference; this makes comparison analysis difficult and sometimes
impossible.
Recidivism
The literature on recidivism goes back at least as far as 1928
(Burgess, 1968). Since then, numerous and diverse studies have been
done. More recently, some researchers have used the meta-analytic
approach that combines several studies, making the review task more
manageable. The focus here is to review studies that identify predictors
of recidivism. Sexual offense is sometimes seen as a special case of
violence, as is spouse abuse. Some studies of these two specialized types
of recidivism are also reviewed.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Recidivism is the word generally used in the literature to refer to
violent and nonviolent reoffenses among those already identified and
adjudicated as offenders. The word comes from the Latin recidere, to fall
back. A forensic recidivist is one who falls back or relapses after being
released from custody for having committed a crime. Instead of rehabilitat
ing, the person relapses or falls back into former behavior patterns and
commits more crimes.
Problems With Definitions
and Data Collection
Older recidivism studies encountered problems with inconsistent
definitions of recidivism. Some studies used rearrest counts as the
measure of recidivism, others used reconviction as the recidivism measure,
and still others used reincarceration as the recidivism measure. These
problems were largely but not completely resolved with the implementation
of the Bureau of Justice Statistics (BJS). While no system is perfect, the
BJS database and the resulting BJS-generated reports present useful
information on a large scale not possible before. The BJS published two
large recidivism reports, one for the year 1983 (Beck & Shipley, 1989) and
another for the year 1994 (Langan & Levin, 2002). Each report looked at
multiple state records and tracked parolees and probationers for 3 years
after their release from custody.
Still, problems remain with incomplete criminal justice data due to
many factors. The majority of criminal court personnel and prosecutors are
from county agencies, which must transmit the outcome of court actions,
thus adding another place for human error to contaminate the accuracy of
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the data on recidivism. Correctional data are generated at both the county
(jail) and state (prison) levels. Then there are criminal activities by one
criminal in several different states, which causes yet another problem for
complete information gathering. Fortunately, the BJS currently keeps track
of all of these contingencies. The BJS maintains arrest, conviction, and
incarceration records for all states, as well as a host of other data reported
annually.
There is a long history of research in the literature focused on the
discovery of predictors of reoffense or recidivism among those persons
already identified as criminal offenders. By far the majority of this research
was done in North America and Europe. The populations studied varied,
including adult felons released from prison, insanity acquittees released
from psychiatric institutions, adolescent offenders, substance abusers, and
domestic violence perpetrators.
If the predictors of recidivism were dependent on such variations in
population characteristics, sensible statements about the predictors
of recidivism would be impossible. Fortunately, however, the same
personal characteristics are related to recidivism regardless of the
population of offenders examined. (Quinsey, Harris, Rice, &
Cormier, 1998, p. 28)
This literature review of recidivism examines studies of several
populations: (a) general male prison/jail population; (b) spouse abusers;
(c) adult male sex offenders; (d) drug, alcohol, and dually diagnosed
offenders; (e) juvenile offenders; (f) foreign offenders, (g) civil committed
and insane psychiatric offenders; and (h) the population of the civil psychi
atric MacArthur Violence Risk Assessment Study. Many predictors of
reoffense are not population specific, which this review demonstrates.
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The U.S. Department of Justice keeps records and produces reports
regarding recidivism (Beck & Shipley, 1989; Langan & Levin, 2002). The
recidivism data from these two report years, 1983 and 1994, are compared
and summarized herein.
General Male Mail Prison/Jail Population
The 1983 BJS Report on Recidivism
According to the 1983 BJS report on recidivism (Beck & Shipley,
1989), 108,580 persons released from prison in 11 states in 1983, which
represents over half of all released state prisoners that year. The rearrest
rate was an estimated 62.5% for felony or serious misdemeanors within 3
years. There were 46.8% prisoners reconvicted and 41.4% returned to
custody of jail or prison. Each person had an average of over 12 prior
offenses, and 2 out of 3 had at least one arrest for violence and a prior jail
or prison sentence for a prior offense. By the end of the 3 years, those
who were rearrested averaged 4.8 new charges. Nearly 23% of those
rearrested were placed in custody for a new violent offense.
The reoffense rate was highest in the 1 st year: 1 out of 4 of those
released were rearrested in the first 6 months, and 2 of 5 were rearrested in
the first 12 months. The rearrests were higher among men, Blacks,
Hispanics, and persons who had not completed high school than among
women, Whites, non-Hispanics, and high school graduates. The younger
the person’s age at the time of prison release, the higher the rate of
rearrest. Seventy-four percent of those with 11 or more prior arrests were
rearrested, compared to 38% of the first-time offenders.
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Prisoners ages 18-24 with 11 or more previous arrests accounted for
an estimated 94% of the rearrests within 3 years of release from custody.
The rearrest rate for property offenders was 68%; for violent offenders, the
rate was 54.6%; for public order offenses, the rate was 54.6%; and for drug
offenders, the rate was 50.4%. About 40% of those released had previ
ously escaped or been absent from custody without leave. About 73% of
these same prisoners were rearrested within the 3-year time frame.
Released prisoners were often rearrested for the same type of
offense for which they had just been imprisoned. Rapists were 10.5 times
more likely than nonrapists to be arrested again for rape. Murderers were
5 times more likely than others to be arrested again for homicide. About
6.6% of the released murderers were rearrested for murder. Almost 1 in 3
released violent offenders and 1 in 5 nonviolent property offenders were
rearrested for a violent crime. Of released burglars, 31.9% were rearrested
for burglary; of drug offenders, 24.8% were rearrested for a drug offense;
and of robbers, 19.6% were rearrested for robbery.
The 11 states chosen for this analysis (California, Florida, Illinois,
Michigan, Minnesota, New Jersey, New York, North Carolina, Ohio,
Oregon, and Texas) and the Federal Bureau of Investigation provided the
criminal history data for this recidivism study. These reports included
information on arrests, prosecutions, court appearances, and postsentenc-
ing statutes such as incarceration, probation and parole. Additional demo
graphic characteristics and postrelease supervision status were provided by
the department of corrections in each state as part of the BJS’s annual
Nation Corrections Reporting Program (NCRP).
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The original sample of released prisoners included 16,355 released
prisoners minus 159 prisoners who had died during the follow-up period.
Complete records were found for 90% of the original sample. Most of the
incomplete records did not have an FBI identification number needed to
match with their corrections records. Without an FBI number, rap sheets
could not be obtained. There was no systematic difference between those
with and those without rap sheets.
This BJS report of the prisoners released in 1983 lists eight risk
factors that give independent net effects on the odds of rearrest. In order of
adding the most effect to adding the least effect, they were (a) age when
released (younger more at risk); (b) number of prior arrests; (c) prior escape
or probation or parole; (d) most serious offense of robbery, burglary, or theft
more at risk, others less at risk; (e) prior incarceration more at risk, no prior
incarceration less at risk; (f) age at first arrest (if 17 or less, then more at
risk); (g) prior arrest for violent offense (if yes, more at risk; if no, less at
risk); and (h) prior arrest for a drug offense (if yes, then more at risk; if no,
then less at risk).
Comparison of Recidivism Rates for Prisoners
Released in 1983 and 1994
The 1994 BJS recidivism report (Langan & Levin, 2002) did not
present risk factors on the odds of rearrest. However, the report presented
comparisons to the 1983 report. Both of the reports included “ out-of-state”
rearrests for those who had been released from prison during that year.
This out-of-state record matching was done for the first time in the 1983
data analysis.
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The rearrest rate rose significantly in 1994. In 1983, the rate of
rearrest was 62.5% of 108,580 state prisoners released from prisons.
In 1994, the rate of rearrest was 67.5 % of 272,111 released from state
prisons. There was also a significant increase from 1983 to 1994 in the
rearrest rate of several other crime categories: released property offenders
from 68.1% to 73.8%, released drug offenders from 50.4% to 66.7%, and
released public-order offenders from 54.6% to 62.2%. However, there was
no significant difference in the rates of rearrest for violent offenders, which
was from 59.6% to 61.7%.
The overall reconviction rate did not significantly change from 1983
to 1994. The rates were 46.8% and 46.9%, respectively. Also, the recon
viction rates were about the same for violent offenders (41.9% and 39.9%),
released property offenders (53.0% and 53.4%), and released public order
offenders (41.5% and 42.0%). There was a significant change in recon
viction of drug offenders, whose rate rose from 35.3% in 1983 to 47.0% in
1994. The reconviction was not necessarily for another drug offense.
In summary, these large-scale recidivism studies highlight recidivism
as rearrest rates in the general population. High risk factors for rearrest
include younger age, male gender, incomplete high school education, and
Black or Hispanic ethnicity. While these two large-scale recidivism studies
were well designed and executed, they did not examine personality factors
and mental health issues. Additional research must be reviewed to
examine mental health issues. For more detail on the methodology used in
each study, see Beck and Shipley (1989) and Langan and Levin (2002).
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Meta-Analytic Studies of Recidivism
The next two studies employ meta-analysis to review and identify
predictors of recidivism. Meta-analysis has certain advantages over
narrative reviews (Cooper & Hedges, 1994). Meta-analysis uses the results
from independent studies and translates them into a common statistic that
is used to compare outcomes across studies. The common statistic, or
effect size, is an estimate of the magnitude of the relationship between two
variables (Bonta, Law, & Hanson, 1998).
Gendreau, Little, and Goggin (1996) used 131 studies that produced
1,141 effect sizes with recidivism. Their strongest predictor domains were
(a) criminogenic needs (such as companions, interpersonal conflict, sub
stance abuse and antisocial personality factors), (b) criminal history/history
of antisocial behaviors, (c) social achievement, and (d) age/gender/race
and family factors. Predictors that were less robust included intellectual
functioning, personal distress factors, and socioeconomic status in the
family of origin. Static and dynamic domains both performed well. The
Level of Service Inventory, Revised (LSI-R; Andrews & Bonta, 1995), which
assesses criminogenic needs, was found to be the most useful actuarial
prediction measure.
The authors of this study concluded that both static and dynamic
measures are useful in the prediction assessment process. They also sug
gested that, when using actuarial measures, the protocols should include
reliable information on early family life and more recent and current social
adjustment. Dynamic risk factors must be included and reassessed over
time; these especially include criminogenic needs, such as companions,
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
interpersonal conflict, substance abuse, and antisocial personality factors.
They recommended using the LSI-R to assess criminogenic needs and the
Psychopathy Checklist Revised (PCL-R) as a general measure of risk.
Using 64 unique samples in their meta-analysis, Bonta et al. (1998)
calculated 35 predictors of general recidivism and 27 predictors of violent
recidivism. For general recidivism, these researchers found in the demo
graphic domain that variables younger age, male gender, and single marital
status predicted recidivism. Minority race and lower social class were not
related to recidivism. In the criminal history domain variables such as
juvenile delinquency and the number of prior convictions predicted recidiv
ism. However, the seriousness of a violent offense, such as homicide or a
sexual offense, was inversely related to general recidivism but the use of a
weapon was positively associated with general recidivism. A history of
violent behavior correlated positively with general offence. The deviant
lifestyle domain variables showed only moderate relationship with recidiv
ism. In this domain the significant predictors of recidivism were family
dysfunction, substance abuse, and poor relationships. The substance
abuse predictive validity came mostly from drug abuse rather than from
alcohol abuse. Education and employment issues were not significant.
The important predictors in the clinical domain included antisocial person
ality diagnosis and a history of psychiatric admissions.
In the meta-analysis by Bonta et al. (1998) of the studies comparing
mentally disordered offenders with nondisordered offenders, the mentally
disordered showed less recidivism. Clinical and objective measures of risk
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
both predicted outcome, but the objective measures rated as better pre
dictors and had less variability in findings.
Bonta et al. (1998) identified 27 predictors of violent recidivism. In
the demographic domain younger age, male gender, and single marital
status predicted violence. In the criminal history domain all of the variables
were predictive of violence except an index offense of violence or sexual
offense. A violent history was more predictive than a violent index offense.
In the deviant lifestyle domain the variable of having a poor work adjust
ment was the best predictor of violent recidivism. Within the clinical domain
antisocial personality was the most potent predictor over any other diag
nostic category. The objective risk assessments were better predictors
than were clinical judgments.
The authors concluded that there were four main predictors for
recidivism for general and violent offenses: adult criminal history, juvenile
delinquency, substance abuse, and antisocial personality. The authors also
stated that predictors of recidivism were almost identical for both mentally
disordered and nondisordered offenders. They concluded that most of the
variables in clinically oriented assessments were either unrelated or
inversely related to recidivism.
Bonta et al. (1998) concluded that several mental disorders such as
psychosis and schizophrenia were inversely related to general and violent
recidivism, while most mood disorders, such as depression, evidenced no
relationship to recidivism. They postulated that these often-assessed
clinical variables showed no utility in prediction because of the episodic
nature of the disorders.
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In summary, these meta-analytic studies identified a large number
of predictors of reoffense. The groupings of predictors included
demographic, criminal history, deviant lifestyle, and clinical predictors.
From items in these groupings four main predictors were identified: adult
criminal history, juvenile delinquency, substance abuse, and antisocial
personality. The authors (Bonta et al., 1998) concluded that psychosis and
schizophrenia were inversely related to reoffense.
Spouse Abuse Offenders
A growing literature addresses spouse abuse and violent reassault of
an intimate partner. Saunders (1995) reviewed the predictors of repeated
wife assault and found almost every risk factor of spouse abuse that has
been identified as a static predictor of criminal violence in general (Quinsey
et al., 1998). Empirical recidivism research has also found that many of the
predictive factors of general violence are associated with spousal assault in
community samples (Grann & Wedin, in press; Hanson, Cadsky, Harris, &
Lalonde, 1997).
Quinsy et al. (1998) and others, including proponents of the PCL-R
(Hare, 1991), have reported that spousal violence and violence in general in
all recidivist populations are related to the same personal characteristics
(Bonta etal., 1998).
Hilton, Harris, and Rice (2001) studied 508 offenders, 88 of whom
were wife assaulters. The wife assaulters scored much lower on the PCL-
R, with a mean score of 11.92 versus a mean score of 18.24 for the other
offenders. They were also much more likely to be rated as suicidal or
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
jealous at the time of the index offense. PCL-R scores were the best
predictor of violet reoffense in the whole sample and with the spouse
abusers who reassaulted.
Gondolf and White (2001) investigated batterer recidivism by
broadening the concept of psychopathy to include dimensions of personality
from scores on the Millon Clinical Multiaxial Inventory, Version III (MCMI-
III). This instrument is a 175-item self-report personality inventory. It yields
24 clinical scales and four validity scales. These researchers found that the
MCMI-III identified repeat assaulters as having psychopathic tendencies.
The MCMI-III further identified a small portion (11%) of the sample popula
tion that exhibited primary psychopathy—the classically known, cold
blooded type of most concern. The MCMI-III classified nearly 60% of the
wife assaulters as subclinical or as having low levels of personality
problems. The MCMI-III scores showed about 50% of the profiles in a
narcissistic continuum and the other 50% in an avoidant continuum. The
authors concluded that it is difficult to typify the personalities of these high-
risk violent cases. In summary, these studies report that psychopathy is a
predictor of reoffense with spouse abusers.
Sex Offender Studies
Hanson and Harris (2000) reviewed the files of 208 sex offender
recidivists and 201 sex offender nonrecidivists. They found that the recidiv
ists were generally considered to have poor social supports, attitudes of
sexual tolerant of sexual assault, antisocial lifestyles, poor self-manage
ment strategies, and difficulties in cooperating with supervision. The overall
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
mood was similar in both groups, but recidivists showed increased anger
and distress immediately before reoffending. The supervising officers’ case
notes found (to a lesser extent) that these findings could not be attributed
solely to recall bias. The objective or static risk scales used in this study
were the Violence Risk Appraisal Guide (VRAG), the Statistical Information
on Recidivism (SIR) scale, the Rapid Risk Assessment for Sexual Offense
Recidivism (RRASOR; Hanson, 1997), and the PCL-R (Hare, 1991).
The significant predictors of both stable and acute risks were
(a) substance abuse, (b) use of antiandrogens, (c) low remorse/victim
blaming, (d) seeing self as low risk to recidivate, (e) access to victims,
and (f) a dirty, smelly presentation. There was no significance in the age
at index offense, but those who reoffended tended to be younger (mean
age of 36.3 years versus 39.1 years). The recidivists’ victims tended more
to be strangers than anyone related.
The best three predictors of the static risk factors were the VRAG,
sexual deviance, and lower IQ. The three best predictors of the stable
dynamic risk factors were seeing self as no risk, poor social influences, and
showing sexual entitlement. The three best dynamic risk factors were
access to victims, poor to no cooperation with supervision, and increased
anger. The stable dynamic risk factors were the most predictive of
reoffense. This was partly due to the design of the study that matched
participants on static history factors. These identical dynamic risk factors
were found by research on nonsexual criminals (Quinsey et al., 1998;
Zamble & Quinsey, 1997).
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Hanson and Bussiere (1998) conducted a meta-analytic review of
61 follow-up studies of sexual offenders’ recidivism on 28,972 sexual
offenders. This study reported eight predictors of sexual offense recidivism:
(a) young age; b) single marital status; (c) antisocial personality disorder;
(d) total number of prior offenses; (e) prior sexual offenses (a higher risk if
victims were strangers, extra-familial, or male victim, if offense occurred at
an early age, or if a variety of sexual offenses); (f) sexual deviance (the
strongest predictor in this meta-analytic study); (g) higher femininity scores
on the MMPI masculinity-femininity scale (Hathaway & McKinley, 1983; but
this was based on only three studies); and (h) failure to complete treatment
(a moderate predictor). None of the other clinical ratings, such as denial or
low treatment motivation, was related to sexual recidivism, except in one
study (Maletzky, 1993).
In this meta-analysis review, Hanson and Bussiere (1998) found that
predictors of nonsexual violent recidivism were the same risk factors com
mon to general criminal populations. These predictors were (a) young age,
(b) unmarried status, (c) minority race, (d) diverse criminal behavior history,
and (e) antisocial or psychopathic personality disorder. The number of prior
sexual offenses was not related to nonsexual violent recidivism. It also
appeared from the analysis that neither sexual deviancy nor subjective
stress was related to nonsexual violent recidivism.
This meta-analysis by Hanson and Bussiere strongly suggests that
sexual offenders differ from nonsexual offenders (Hanson, Scott, & Steffy,
1995); in particular, sexual offenders are more sexually deviant in their
interests than are nonsexual offenders (e.g., interest in children, especially
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
boys). The authors suggested that, to differentiate sexual offenders from
nonsexual offenders, risk assessments should include some means of
measuring sexually deviant interests. Two additional studies (Gretton,
McBride, O’Shaughnessy, & Hare, 1997; Rice & Harris, 1997) found
increased risk for sexual offending using the combination assessment for
sexual deviancy and for psychopathy.
In summary, when examining the sex offender for risk of reoffense,
psychopathy is still a major risk factor, but sexual deviancy must also be
assessed.
Substance Abuse and Dual Diagnosis Offenders
Prison Samples
Employing a sample of 728 Chicago inmates, Abram (1989) found a
significant relationship between drug abuse disorders and violence. How
ever, Abram found no significant relationship between alcohol abuse and
violence. Abram found that prior violence was the best predictor of
reoffense and violence.
In Ontario, Canada, Loza (1993a, 1993b) looked at cohort treatment
needs with the Level of Service Inventory (LSI) and the PCL-R factors 1
and 2. (Factor 1 refers to personality and affective traits in a person and
factor 2 refers to the life style and behavior habits or patterns characteristic
of psychopaths.) Loza found that 76.69% of this group abused alcohol,
illicit drugs, or both. The group that abused both alcohol and drugs had a
31.29% reoffense of violent offense and obtained the highest scores on
both the LSI and the PCL-R. Next, the “drug-only” abusing group had a
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30.06% reoffense of violent and the next highest scores for needs on both
instruments. Alcohol-only abusers had a 15.40% reoffense rate and lower
needs scores. Suggestions by the author included that both Axis I drug and
alcohol issues and Axis II criminogenic diagnoses should be addressed in
follow-up of these parolees.
In San Bernardino County, California (Dalton, 2000), a group of 518
inmates chosen for their substance abuse diagnosis was part of a study to
examine the effects of substance abuse interventions in the jail setting. Of
the 255 substance-abusing males, 132 were in an intervention group and
123 were in a comparison group. There were 139 females in another
intervention group, with a comparison group of 124. Almost 75% of the
comparison group (no treatment and no interventions) were re-arrested
within an 18-month follow-up period. In contrast, about 55% of the interven
tion group were re-arrested during the same time period. Among females,
there were 16% more recidivists in the comparison group than the inter
vention group. Among males, 62.6% were recidivists in the comparison
group and 35.4% were re-arrested during the 18-month follow-up period.
Prior criminal history was clearly the best predictor of recidivating.
In a federal prison, 115 inmates participated in a study by White,
Ackerman, and Caraveo (2001). The researchers found that self-reports of
alcohol abuse were more likely to show Antisocial Personality Disorder,
Anxiety, and drug misuse. In this study 24% showed evidence of primary
psychopathy, 61% tested positive for alcoholism, 32% had been polydrug
abusers in the previous 12 months, 22% marijuana exclusively, and 10%
used cocaine, heroin, or amphetamines exclusively. Those whose alcohol
21
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and drug screens were positive were 3 times more likely to have been
violent against a domestic partner during the prior 12 months.
White et al. (2001) cautioned against making generalizations from
their study, citing the need for studies with larger samples. They also noted
that the PCL-R should be used to identify psychopathy as a control for
spurious contributors to arrest and crime. They stressed the need for
alcohol and drug treatment in low-security prisons where inmates serve
relatively short sentences and then return to the community and their
families. These inmate studies point to criminal history and relapse to drug
or alcohol abuse as risk factors for reoffense.
Community Studies
In London, Scott et al. (1998) examined 92 aggressive psychiatric
patients (65 who had psychosis only and 27 who were dually diagnosed
with psychotic illness and substance abuse). Recent aggressive behavior
was analyzed and did not reach significance, although the dually diagnosed
showed a trend of reporting offences more frequently. There was a highly
significant association among aggressive/hostile behavior, lifetime history
of offence, and recent history of assault. Although a direct causal link
between dual diagnosis and aggression has not been established, it is
important to investigate psychiatric patients for dual diagnosis.
Fulwiler, Grossman, Forbes, and Ruthazer (1997) compared 37
patients with a history of violence in the community with 27 patients without
such a history. The only significant difference between those with a history
of violence and those without a violence history involved alcohol or drug
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
use. Fulwiler et al. concluded that very early onset of substance abuse
among those with major mental illness was associated with the greatest risk
of community violence. The authors suggested that some of the determin
ants of violence in this sample might precede the onset of adult mental
illness. In this study only 7% of the patients with major mental illness only
had a history of violence, but 73% of the patients who had both mental
illness and substance abuse had a history of violence. In this sample both
the rate of previous violence and the rate of substance abuse were high.
Participants were from low-income inner-city neighborhoods in Boston. The
authors found the following meaningful correlations to violence: (a) sub
stance abuse and noncompliance to treatment, (b) substance abuse and
being more severely ill or disturbed (may be related to unremitting signs of
active psychosis), (c) substance abuse and poor insight, (d) substance
abuse and early onset of drug or alcohol abuse, and (e) early onset of
childhood conduct disorder.
Drake, Osher, and Wallach (1989) studied 115 schizophrenic adults
ages 20-65 years who were discharged from state hospitals and partici
pated in urban aftercare programs. Over a 6-month period, 45% used
alcohol and 22% were clearly abusing alcohol. Alcohol use was associated
with younger age, male gender, street drug use, medication noncompli
ance, psychosocial problems, increases in symptomatology, chronic
medical problems, and a higher rate of rehospitalization. Drake and
colleagues concluded that even a mild use of alcohol might be problematic
for schizophrenic patients. Notably missing was information on multiple
diagnoses, psychopathy, or antisocial personality disorder.
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In a 22-year follow-up study, Yesavage, Benezech, Larrieu-Arguille,
and Bourgeois (1986) examined mentally ill patients judged not responsible
for an act of crime. On their first admission into the hospital, 42% had
committed crimes of murder and assault. During the 22-year follow-up, the
rate of violent crime increased to 46% at readmission. The subjects’ data
were analyzed by diagnostic category. Personality disordered patients’
rates of violence increased from 28% at initial admission to 38% at
readmission. Mental retardation patients’ rates of violence increased from
46% at initial admission to 59% at readmission. Schizophrenic and para
noid psychosis rates of violence remained the same: 22% and 43%,
respectively. Alcoholism rates of violence decreased from 64% to 57%.
The total percentage of violent patients on first admission associated with
alcohol was 19%, increasing to 35% for recidivists. However, the authors
conclude that alcohol abuse and mental illness lead to violence.
Wessely (1998) retrospectively examined 20 years of first hospital
admissions for schizophrenics. He commented that not all schizophrenics
are admitted into hospitals, which is a reason for bias in this type of study.
He stated that the decision to admit is associated with the presence of
violent behavior and/or police involvement (Castle, Phelan, Wessely, &
Murray, 1994). This means that some bias is introduced in studies of crime
and schizophrenia that exclude this small number. Also, during the time of
this study, only one homicide was recorded. Thus, difficulties result with
analysis and prediction of this very low rate of violence. In this study, the
predictors of convictions were previous conviction, ethnicity (African
American), substance abuse, and schizophrenia.
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Swartz et al. (1998) interviewed 331 severely mentally ill patients.
They examined medication noncompliance and substance abuse as
correlated to (or predictors of) violence. Interviews were also done with
family and other informants to obtain information about the 331 patients
during the 4 months previous to hospitalization. Those with substance
abuse were found to be 2 times more likely to be violent and those who
were noncompliant with medications were 2.29 times more likely to be
violent than the patients who did not abuse substances and were compliant
with medications. The authors concluded that those who abused alcohol
and illicit drugs were at high risk to be violent.
In a study of 1,136 patients Steadman et al. (1998) found substance
abuse to be strongly associated with violent behavior. The authors com
pared violence rates among the following groups: (a) those with major
mental illness and no substance abuse, (b) those with major mental illness
and substance abuse, (c) those with other mental illness and no substance
abuse, (d) those with other mental illness and substance abuse, (e) those
with no mental illness and no substance abuse, and (f) those with no mental
illness and substance abuse. There was no difference between the groups
with no substance abuse. The major mental illness and substance abuse
group showed a 31.1% prevalence of violence. The other mental illness
group’s prevalence rate was 43.0%. The rate for major mental illness and
no substance abuse was 17.9%. The authors concluded that substance
abuse increases risk of violence among those with mental illness.
Raesaenen et al. (1998) studied a birth cohort of 11,017 persons up
to age 27 in Finland. Finland maintains regular and complete psychiatric
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and medical records. Over 90% of violent crimes in Finland are solved
(compared to about 70% of violent crimes solved in the United States).
These Finnish scholars found that schizophrenic men who abused alcohol
were 25.2 times more likely to commit violent crime than mentally healthy
men. The risk of committing violent crime for nonalcoholic patients with
schizophrenia was 3.6 times, and the risk for other patients was 7.7 times
that for healthy men. None of the patients with schizophrenia (who did not
abuse alcohol) were recidivists. The risk for reoffense for alcoholic subjects
with schizophrenia was 9.5-fold. The results appeared clear that comorbid
alcohol abuse with schizophrenia is a high risk factor for violence.
The literature cited in the immediately preceding paragraphs reports
that risk factors for reoffense among the mentally ill include criminal history,
drug and alcohol abuse, early life onset to drug abuse, and early life onset
to conduct disorder.
Adolescent Offenders
Dean, Brame, and Piquero (1996) followed 848 teens, age 16 and
older, for just under 6 years (1988-1994). The study focused on the
correlates of offending persistence in two categories: early age and later
age at first adjudication. Differences were found when the starting age for
the late adjudication group was set for age 12. When the starting age was
increased, the differences disappeared. Typology theorists, such as Moffit
(1993) and Patterson (Patterson & Yoerger, 1993), have stated that there
are two different kinds of offenders. The study by Dean et al. (1996)
supports that early starts predicted longer and more frequent reoffending.
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Child abuse at a young age correlated with the early start of offending but
was not significant for late starters.
In summary, the adolescent offenders show higher risk for reoffense
by criminal history and early life problems. These are factors in the assess
ment of psychopathy and are in the PCL:SV.
Foreign Offender Populations
Finland Offenders
Koskinen et al. (2001) studied a 1966 birth population of 5,589 males
in northern Finland. The results showed that the risk of reoffending was
almost 8 times the baseline if the child was born and raised by a single
mother until age 14. The probability was 5 times the baseline if the mother
married later during the boy’s childhood, was 2 times the baseline if the
parents were at first together and then separated. This study highlights the
enormous impact of parental instability on early childhood and later adult
behaviors.
Canadian Offenders
Porter, Birt, and Boer (2001) studied 317 Canadian federal offenders
using the PCL-R (Hare, 1991). There were 224 low scorers (scores below
30) and 93 who scored in the psychopathic range (over 30) of the PCL-R.
High scorers in the psychopathic range committed more violent and non
violent crimes than the low scorers for about 30 years, from late adoles
cence into their 40s. The numbers of nonviolent crimes decreased much
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
after age 30, but violent crimes decreased and then increased again in the
late 30s before significant reduction was evident.
Nordic Offenders
Pinard and Pagani (2001) reviewed studies, mostly in Nordic
countries, of the association of mental disorder and recidivism. They found
that early-onset alcohol abuse (type 2 alcoholism) and schizophrenia
contributed to recidivistic violent behavior.
These Canadian and Scandinavian studies support as high-risk
predictors of reoffense the following: high scores on psychopathy, early
onset of criminal behavior, and early onset of alcoholism.
Civilly Committed and Insane Offenders
Sreenivasan, Kirkish, and Eth (1997) studied 109 males from two
sites: a sample of 66 consecutive involuntary civilly committed (WIC 5150)
patients from the Veterans Administration Hospital admissions and 43
Patton State Hospital patients found not guilty by reason of insanity (PC
1026). All were males, ages 19 to 49 years. The total sample was divided
by those with two or more lifetime acts of physical harm to a person, called
high violent (HV), versus nonviolent acts such as verbal threats, vandalism,
drunkenness, or theft, called low violent (LV).
Sreenivasan et al. (1997) found three routes of pathology associated
with violent recidivism: (a) dual diagnosis of psychotic disorder and sub
stance abuse; (b) cognitive inflexibility as measured by the Wisconsin Card
Sorting Test, scoring in the moderate to severe range in perseveration; and
(c) psychopathic traits as assessed by PCL-R scores in the moderate range
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(20-29). PCL-R scores, combined with the Block Design, Trail Making B of
the Wechsler Adult Intelligence Scale-Revised (WAIS-R), correctly identi
fied 84.2% of the civilly committed HV group and 88.96% of the combined
LV group. The authors concluded that violence treatment models should
include assessment of three violence trigger controls: reality-testing
capacity, cognitive flexibility, and moral controls.
The MacArthur Violence Risk Assessment Study (Monahan et al.,
2001; Steadman et al., 1998) found psychopathy as measured by the
PCL:SV to be a strong predictor of violence among a sample of 1,136 civil
psychiatric patients followed for 12 months (Skeem & Mulvey, 2001a,
2001 b). The domains assessed for covariants of violence in this study were
(a) criminal and violence history; (b) substance use diagnosis; (c) person
ality disorder diagnoses; (d) anger, impulsivity, and antagonism; and
(e) demographic characteristics. The predictors of violence from these
five domains were (a) gender, (b) early childhood experiences, (c) neigh
borhood risk factors, (d) diagnosis, (e) psychopathy assessed by the
PCL:SV, (f) delusions, (g) hallucinations, (h) violent thoughts, and (i) anger.
A summary of each of these predictors identified in the MacArthur Violence
Risk Assessment Study follows below.
Regarding the predictor gender and violence, Monahan et al. (2001)
found that (a) men were no more likely than women to be violent in the 12
month follow-up period, and (b) prior violence and criminality correlated with
the pattern of post-discharge violence for both males and females—for both
African Americans and Whites—with this exception: For females with prior
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
arrest for crimes against a person versus those with no arrest record, the
difference was not significant.
Regarding early childhood experiences and violence, physical abuse
as a child correlated with later acts of violence. Sexual abuse as a child did
not correlate with later acts of violence. Drug abuse by the father correlated
with later acts of violence. Having a father in the home until age 15
reduced later acts of violence (explained as drug-abusing fathers not being
in the home as much as nonabusing fathers).
Regarding neighborhood risk factors and violence, concentrated
poverty in the neighborhood correlated with violence. Socioeconomic
status was less correlated with violence than was neighborhood poverty.
Regarding diagnosis and violence, three diagnostic groups were
analyzed for correlates to violence: (a) major mental disorder— psychosis,
depression—and no co-occurring substance abuse, (b) major mental
disorder with co-occurring substance abuse, and (c) other mental disorder
with substance abuse.
Regarding psychopathy as assessed by the PCL:SV, psychopathy
correlated with violence. However, in civil psychiatric patients the PCL:SV
may be assessing emotional detachment as much as antisocial behavior.
The authors suggested caution when assessing this population, with this
clarification. They stated that total scores reflecting high scores on anti
social behavior may show a higher risk for violence than similar total scores
that are based on emotional detachment (Rogers, 1995).
Regarding delusions and violence, the findings were contrary to con
ventional and popular wisdom. The presence of delusions was associated
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
with less violence than the absence of delusions over the entire 12-month
follow-up period. Those with delusions showed a violence rate of 22.8%
rate; those with no delusions showed a violence rate of 29.4%. Those with
threat/control override (TCO) delusions had a 19.7% rate of violence; those
without TCO delusions showed a 29.4% rate of violence.
Regarding hallucinations and violence, those with hallucinations
were more likely to be violent. Those whose command hallucinations told
them to be violent were significantly more likely to be violent than those
without command hallucinations, 44.6% and 23.9%, respectively.
Regarding violent thoughts, if a patient reported violent thoughts
during a recent treatment assessment in the hospital, there was a greater
likelihood that he or she would act violently upon discharge.
Regarding anger and violence, patients with high anger scores on
the Novaco Anger Scale (NAS; Novaco, 1994) were twice as likely as those
with low anger scores to engage in violence after release in the 12-month
follow-up period.
The researchers of the MacArthur Violence Risk Assessment Study
(Monahan et al., 2001) concluded that violence risk assessment remains a
complex task requiring an interactional approach to the risk assessment.
These authors proposed a decision tree based on the research findings and
that patients be assigned to one of 11 separate risk groups, depending on
the decision tree sort. The risk of violence ranges from low to high in each
of the groups and goes from a low, low score of 0.5 to a high, high score of
63.8.
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Summary of General Male Recidivism Studies
From the studies reviewed there emerges a core set of predictors of
recidivism. This set includes early childhood events of physical abuse,
family disruption, and early onset of criminal behavior; younger age of
arrest and release from custody; more variation in types of crime; more
prior crime, including prior violent crime; attitudes permissive or condoning
of deviance; dual diagnosis of substance abuse and mental disorder; and
poor social adjustment in the community.
Generally, serious mental illness is associated with increased risk of
criminal and violent behavior. The greatest risk among the mentally ill
offenders is a co-occurring substance abuse problem and antisocial/or
psychopathic personality.
The Three Hypothesized Predictors
and Offender Studies
The three hypothesized predictors have some supporting research
that they are predictors of criminal and violent offense with mentally dis
ordered offenders. Each predictor is reviewed briefly.
Psychopathy Studies Using the PCL.SV
The rationale for examining the relationship of psychopathy to
reoffense and violent behavior in this study is that many of the studies
linking psychopathy to reoffense have been done with prison and hospital
populations rather than with psychiatric outpatient offenders on conditional
release. However, Harris et al. (1993) found that psychopathy, measured
by the 20-item PCL-R, was the best single predictor of violence in groups of
schizophrenic and personality-disordered individuals. The study by Harris
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
et al. is among those that have established the utility of the psychopathy
construct with the mentally disordered offender population. Many
researchers have found psychopathy to be a robust risk factor for future
violence in a variety of populations (Forth, Hart, & Hare, 1990; Hill, Rogers,
& Bickford, 1996; Quinsey, Rice, & Harris, 1995; Serin, 1991,1996; Serin &
Amos, 1995). In an analysis of 18 studies Salekin, Rogers, and Sewell
(1996) found a large effect size (d= .79) between psychopathy and
violence (and other forms of antisocial behavior).
The debate has continued surrounding the etiology of those charac
teristics labeled psychopathy, antisocial personality disorder, and socio-
pathy. There seems to be consensus concerning the disorder’s core
“ affective, interpersonal and behavioral attributes” (Hare, 1996, p. 27).
There also is significant agreement on the descriptive features of this dis
order among many professional disciplines and nonprofessionals, including
criminal justice personnel, psychologists, psychiatrists, and laypersons. As
found in many studies of recidivism, the behaviors that correlate with and
are said to predict recidivism find a place in the PCL-R and PCL:SV
assessment on factor 2, deviant behaviors. The correlates described by
items on factor 1 of the PCL, affective and interpersonal behavior patterns,
contribute additional explanations or predictions of reoffense. Some
relevant empirical studies show this.
General Male Offender Studies of Psychopathy
Hemphill, Hare, and Wong (1998) conducted a meta-analytic study
of 10 previously reported analyses of psychopathy and recidivism (N =
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1,275). They found the average weighted correlation of psychopathy
scores to general recidivism to be .27 and to violent recidivism to be .27.
General recidivism was more strongly correlated with factor 2 (a measure of
social deviance) than with factor 1 (a measure of affective and interpersonal
facets of psychopathy). However, violent recidivism did not more strongly
correlate with either factor 1 or factor 2; the factors contributed equally to
the correlation.
In hierarchical regression studies where demographic (age) and
criminal history variables were removed from the analysis, the PCL-R pre
dicted recidivism beyond the previously entered criminal history variables
(Harris, Rice, & Cormier, 1991; Hart, Kropp, & Hare, 1988; Heilbrun et al.,
1998; Rice, Harris, & Quinsey, 1990; Ross, Hodgins, &Cote, 1992).
Other studies found that the PCL-R predicted recidivism in male
federal offenders (Hemphill, Templeman, Wong, & Hare, 1998; Zamble &
Palmer, 1996), adolescent offenders (Gretton, 1997; Gretton, McBride, &
Hare, 1995; Gretton et al., 1997; Toupin, Mercier, Dery, Cote, & Hodgins,
1996), civil psychiatric patients (Douglas, Ogloff, & Nicholls, 1997), black
offenders (Cooke, Kosson, & Michie, 2001; Kosson, Smith, & Newman,
1990), sex offenders (Rice & Harris, 1997), and forensic psychiatric patients
(Hill et al., 1996; Rice & Harris, 1992; Wintrup, Coles, Hart, & Webster,
1994). Correlations between recidivism and PCL-R scores in other studies
have been similar to those found in the review by Hemphill, Hare, et al.
(1998).
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Sex Offenders and Psychopathy
Hemphill, Hare, et al. (1998) found that the PCL-R scores were
among the best predictors of violent and sexual recidivism. Using relative
risk statistics, these authors found that psychopaths had general recidivism
rates 3 times those of nonpsychopaths. They found psychopaths to be at 4
times greater risk for violence than nonpsychopaths.
Schizophrenic Offenders and Psychopathy
Nolan, Volavka, Mohr, and Czobor (1999) reviewed hospital and
arrest records of 26 violent schizophrenic/schizoaffective patients and
compared them with records of a matching group of 25 nonviolent patients.
They found that the mean psychopathy scores of violent patients were
higher than those of nonviolent patients. Nineteen percent of the violent
patients (n = 5) had scores exceeding the cutoff for psychopathy, and 50%
(n = 13) scored in the possible psychopathic range. All of the nonviolent
patients’ scores were below the cutoff for possible psychopathy. The
authors concluded that schizophrenics who score high on measures of
psychopathy may indicate a subtype of schizophrenia that shows early
symptoms of conduct disorder and persistent violent behavior.
Tengstrom, Grann, L&ngstrom, and Kullgren (2000) found that
psychopathy as measured by PCL-R predicted violent recidivism among
male violent offenders with schizophrenia (N = 202). Psychopathy was
measured with retrospective file-based ratings. The mean follow-up time
was 512 months. Twenty-two percent of the subjects scored greater than
or equal to 26 (cut-off) on the PCL-R. The area under the curve (AUC) of
the receiver operating characteristic of the total score on the PCL-R to
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
predict violent reoffense ranges from .64 to .75, depending on the time
frames.
The literature reviewed in the preceding paragraphs indicates that
psychopathy and the PCLSV strongly correlate with reoffense in schizo
phrenic, sex offender, spouse-abusing, and general male offender popula
tions.
Active Psychosis Studies
The rationale for examining the relationship of active psychosis and
reoffense and violent behaviors in this study is that a growing number of
research investigations have shown evidence of some link between mental
illness and violent crime. The present author’s 12 years of clinical observa
tion with clients on conditional release from state hospitals also supports
the mental illness variable as a predictor of future reoffense. Monahan
(1992) suggested that florid psychotic symptoms are more highly correlated
with violent behavior than is the simple diagnosis of a psychotic disorder.
In their quantitative review, Douglas and Hart (1996) found a
stronger effect during analysis for specific psychotic symptoms versus
gross categories of mental illness. Link and Stueve (1994) and Swanson,
Borum, Swartz, and Monahan (1996) found that psychotic symptoms, which
override one’s sense of safety or well-being (so-called “ threat/control
override” symptoms), were more strongly related to violence than psychotic
symptoms, which do not possess these qualities. De Pauw and Szulecka
(1998) claimed that there was some evidence that patients “ with well
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
developed delusions are more likely to commit violent crimes against
persons than those with chronic undifferentiated psychoses” (p. 91).
Many studies have observed a relationship between psychiatric
symptoms, as measured by the Brief Psychiatric Rating Scale (BPRS;
Overall & Klett, 1962) and violence (Douglas & Hart, 1996). Other clinical
conditions that should be accounted for include such states as sadistic
fantasies (MacCulloch, Snowden, Wood, & Mills, 1983), suicidality
(American Psychiatric Association, 1974; Brent etal., 1994; Hillbrand,
1995), and paranoia, self-aggrandizement, and pathological jealousy.
Homicidal ideation clearly must be noted and investigated.
Linking active psychosis, specifically delusions and hallucinations, to
criminal behavior could be found to be very similar to linking any mental
illness to criminal behavior. To look at delusions and hallucinations more
narrowly defines behavioral symptoms.
Link and Stueve (1994) proposed a model to explain a link between
mental illness and violence. In this model they stated that the behavior of
the psychotic patient is rational, or is predictable, if one first knows what is
the perceived reality of the patient. For example, a patient would be
expected to defend or fight for his or her life if a mortal threat were believed
to be present. When the threat is perceived but not real, it is labeled a
paranoid delusion. An interesting note is that this model emerged from a
study and analysis undertaken to show that no significant link existed
between mental illness and criminal behavior. The authors stated in this
model that a “rationality” or understandable explanation of a violent act
occurs with some psychotic and irrational symptoms.
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The principle of rationality-within-irrationality posits that once one
suspends concern about the irrationality of psychotic symptoms and
accepts that they are experienced as real, violence unfolds in a
rational fashion. By rational we do not mean reasonable or justified
but rather understandable. (Link & Stueve, 1994, p. 143)
For example, when a person fears personal harm or feels threatened by
others due to delusions or hallucinations, violence is more likely to follow.
Link and Stueve (1994) investigated this model using data collected
by Dohrenwend and colleagues, who used the Psychiatric Epidemiology
Research Interview (PERI; Dohrenwend, Shrout, Link, Martin, & Skodal,
1986; Shrout et al., 1988). Link and Stueve developed a TCO scale using 3
of the 13 items of the PERI’s psychotic symptom scale. The other 10 items
were labeled “other psychotic symptoms.” These scales were used to
analyze the responses to three questions posed to three separate popula
tion samples. Each group was matched between psychiatric patients and
nonpatient community participants. The first group answered the question
about “hitting” in the previous month or year, the second group answered
the question about “ fighting” in the previous 5 years, and the third group
answered the question about “using a weapon” in the previous 5 years.
Due to the way in which the above research information was
collected, the three data sets were analyzed separately. The first group,
“hitting,” showed that 12.3% of patients (n = 385) reported hitting in the
previous year, while 5.2% of the nonpatients (n = 365) reported hitting in
the same time period. Of the second group, “ fighters,” 25.7% of the
patients (n = 191) and 15.1% of the nonpatients (n = 185) reported fighting
in the previous 5 years. In the third group, “used a weapon,” 9.7% of the
patients (n = 195) and 2.7% of the nonpatients {n = 186) reported having
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
used a weapon in the previous 5 years. Utilizing further regression analy
sis, the TCO scale remained a significant predictor of violence. This was
true even when the other psychotic symptoms were held constant. The
same pattern of significance for TCO symptoms remained after other
possible determinants of behavior were included in the analysis.
These studies have shown support for a seriously mentally ill patient
to be at higher risk to commit violence than the non-mentally ill. The
“ threat/control override” concept helps to explain the purpose of violence by
those who are experiencing active symptoms of thought-disordered mental
illness.
Substance Abuse and Mental Illness Studies
The relationship between substance abuse and violence in the
general population has long been debated. Substance abuse has been
shown to be a significant factor in reoffense among seriously mentally ill
offenders in the large Epidemiological Catchment Area study (ECA;
Swanson, 1994). Among the mentally ill, there was a 7% absolute risk of
violence in the course of a year. Among single-diagnosis substance
abusers, there was a self-reported prevalence of violence of 55.2%.
Among those with mental illness and substance abuse (dual diagnosis), the
rate was 63.9%.
Substance abuse problems proved a significant predictor of violence
in the study by Harris et al. (1993). In their recent reviews, Klassen and
O’Connor (1994) regarded substance abuse as an important link to violent
behavior. Other research supports the link between substance use and
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
violence in various sample populations (Bartels, Drake, Wallach, &
Freeman, 1991; Blomhoff, Seim, & Friis, 1990; Hodgins, 1990; Hoffman &
Beck, 1985; Steadman et al., 1998; Taylor, 1985).
A growing body of research has established an association of
criminal or violent offense with dual diagnosis of major mental illness
(thought disorders) and substance abuse disorders. Causal links between
the dually diagnosed and criminal behavior have remained elusive,
although they have been frequently hypothesized. The ECA (Robins &
Regier, 1991) showed a significant correlation between the dually diag
nosed (mental illness and substance abuse disorders) and criminal violence
and nonviolent offenses.
Summary of Hypothesized Predictor Literature
More and more empirical studies confirm the relationship between
psychopathy and increased criminal and violent offenses. Separate and
different populations confirm this finding. The meta-analytic reviews cited
lend collective and strong support for psychopathy as a main predictor of
reoffense and violent behavior. Among sex offenders, measures of
psychopathy, combined with a measure of deviant sexual interest, predict
sexual offense. Recent studies that used the PCL-R with schizophrenic
offenders (Nolan et al., 1999; Tengstrom et al., 2000) show psychopathy to
predict violent recidivism.
Active psychosis as a predictor of reoffense has had mixed reviews
in the literature in the past. A growing majority have shown an increased
risk for reoffense and violence when command hallucinations are currently
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
present and then certain delusions are present that are perceived as
threatening (Link & Stueve, 1994). The dual diagnosis studies show that
mental illness combined with substance abuse increases the risk for
reoffense and violence more than 9-fold (Raesaenen et al., 1998).
Theoretical Limitations
Relevant Theories for This Study
Violence research (Mulvey, 1994; Mulvey & Lidz, 1993; Steadman et
al., 1993), as well as research on mental disorder (Steadman et al., 1994)
and on personality dysfunction (Widiger & Trull, 1994), suffers from the lack
of a unifying and encompassing theory. The investigation of violence has
been done from many separate disciplines, including psychology, sociology,
biology, criminology, and anthropology (Roberts, Mock, & Johnstone, 1981;
Tedeschi & Felson, 1994). This has resulted in the forming of many mini
theories and related concepts (Tedeschi & Felson), all without any over
arching theory that would guide researchers in variable selection.
The combining of mental disorder and violence in research further
exacerbates the already very complex task of attempting theoretically
driven research. There is a noted lack of conceptual clarity regarding how
violence and mental disorder may be linked together (Mulvey, 1994).
Mulvey suggested that there is a need to “ develop specific theories about
the mechanisms linking” violence and mental disorder (p. 667).
Violence research, and particularly prediction research, has been
mostly empirically driven. This approach, which takes previous findings,
refines them, and then incorporates them in future research, has generated
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
extensive literature and is still the driving energy of violence and prediction
studies. Mulvey and Lidz (1993) recommended “ greater specificity” in the
gathering of violence data and that as much information as possible be
included in the data collection of violence research. They also cautioned
that “ collecting information on interesting variables without any overarching
theory . . . can sidetrack a field by forcing it to weed out the spurious
relationships” (p. 279).
A summary of this problem was made by Steadman et al. (1994)
regarding theory, variable selection, and violence:
We are aware of the way that textbooks say variable selection is
supposed to be done by deductions from a fully articulated and valid
. . . theory. We are also aware that no such Theory of violence or
mental disorder exists. Nor is it plausible to hope that we will pro
duce a grand Theory in the near future. Therefore, we took a more
eclectic and heuristic approach to variable selection. We looked at
cues that had been validated . . . in the existing research literature.
. . . We looked at factors mentioned in the clinical literature or
emerged from our own clinical experiences, (p. 47)
This very clear statement reveals the status of theoretical approaches to
violence research. It also explicitly states that the empirical approach has
been adopted by most investigators in their search for links between
violence and mental disorder.
Without a specified overarching theory, the door is open to alterna
tive organizing structures. The one chosen as appropriate for this study is
Engel’s (1980) biopsychosocial model. This model acts as an organizing
structure for treatment and research when the combination of mental illness
and violent behavior is examined.
There are studies that support a biological basis for mental illness
(Sheitman, Kinon, Ridgway, & Lieberman, 1998) and violence (Mednick,
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1977; Mednick & Kandel, 1988; Moffitt, 1993; Raine, Brennan, & Mednick,
1994). There are psychological and social theories of both mental disorder
and violence (Akers, 1998; Dilts, 2001).
Dopamine Hypothesis
A prominent biological theory of schizophrenia is that of the dopa
mine hypothesis. This brain chemistry, neurotransmitter-regulating
approach to treating schizophrenia and violent behavior is the front line of
intervention with those suffering from psychotic disturbances. There is no
approved pharmacotherapy for violent behavior separate from medical and
psychiatric illness. When a patient with schizophrenia begins treatment,
antipsychotic medications that regulate the brain chemistry are prescribed
first. After the florid psychotic features are reduced or brought into
remission through neuroleptic medications, then social skills and problem
solving strategies are incorporated into the biopsychosocial treatment
(Kaplan & Sadock, 1996).
Substance abuse research points to some biogenetic precursors to
addictions and abuse. While genetic predispositions have not yet become
alterable biologically, some markers are known. Almost all substance
abuse treatment modalities use psychological awareness and social
interaction retraining in the successful rehabilitation of persons suffering
from substance abuse problems.
The biopsychosocial model or framework is thus a useful and
relevant organizing structure in selecting variables for research about
predicting violence among conditionally released schizophrenic individuals.
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A three-part theory/model is more relevant than a single minitheory
approach to mental illness and violent behavior. The relevance derives
(some) from the empirical studies that support effective or “ what works”
strategies in the treatment of mentally disordered offenders.
Correctional Treatments That Work
The most promising research of “ what works” has been done with
adolescents, as shown by Lipsey’s (1992,1995) meta-analyses of over 400
studies. The more influential moderator variables in Lipsey’s analyses were
behavioral and skills training, multimodal treatments, higher-risk individuals
receiving more frequent services and more structured treatments. Less
effective were treatments such as counseling only, less structured treat
ments, and nonindividualized treatments.
Adult offenders may be subject to different or additional problems
than adolescents. One adult study, the Cognitive Self-Change program in
Vermont, targeted cognitive behaviors. Those who stayed in the program
for 7 or more months showed a significant reduction in the rate of recidiv
ism, compared to those who had no treatment or who dropped out of treat
ment with 6 months or less in the program. Those who were in the program
for 7 or more months showed a 50% reduction in new accusations in the 1 st
year, 40% fewer accusations in the 2nd year, and 39% fewer accusations
by the end of the 3rd year posttreatment completion.
This outcome study supports a cognitive-behavioral treatment
modality and was designed from learning theory approach (Losel, 1995).
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The Biopsychosocial Model and Violence
The theoretical model used in connection with this study is the
biopsychosocial model introduced into medical and psychiatric treatment
literature by Engel (1980). This model is an integration of three broad
theory fields and fits the research questions and data collected for this
study. This model combines or loosely integrates biological, psychological,
and social theories of behavior, which in turn suggest treatments for prob
lem behaviors. While it has been mostly applied in medical and psychiatric
treatment settings, there is a growing research emphasis for examining the
interaction between biopsychological and social risk factors that predispose
individuals to antisocial and violent behavior (Farrington, 1977; Raine,
1993).
The next few paragraphs summarize with supporting research the
usefulness and appropriateness of the biopsychosocial model approach to
understanding the origins of antisocial behavior and violent acts of schizo
phrenic patients and the hypothesized predictors of reoffense.
Biopsychosocial Model and Active Psychosis
Brain imaging technologies document physical and neurological
differences in the brains of schizophrenic patients versus nonschizo
phrenics (Andreason & Black, 2001). The biological theory of schizo
phrenia implies the risk factor of active psychosis and involves the
neuroscience of brain chemistry and neurotransmitters. The primary point
of intervention to control schizophrenic positive, negative, and residual
symptoms has been the use of neuroleptic medications for the past 50
years. Medications that regulate the brain chemistry of patients suffering
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
from this illness have differences in method of action on the brain
chemistry, and individual patients respond differently to the many various
medications. Frequently, “ typical” medications (historically, dopamine-only
antagonists) are introduced first; then, “atypical” medications (medications
combining dopamine and serotonin antagonists) may be substituted,
depending upon the effectiveness of symptom reduction and presence of
negative side effects from each medication along the “ trial and error” course
of neuroleptic treatment (Andreason & Black, 2001).
The original “ typical” neuroleptic medications have regulated
dopamine activity in the brain as dopamine antagonists. Newer “atypical”
neuroleptic medications have added to and refined the dopamine regulation
action with medications that regulate both dopamine and serotonin as well
as other neurotransmitters. Many of these “atypicals” are more effective
and have fewer or no negative side effects for some patients (Fogel,
Schiffer, & Rao, 2000).
Psychosocial evidence for this model comes from studies utilizing
the prevailing ’’stress-vulnerability-protective factors” model of schizo
phrenia, which includes family interaction, emotional expressions, and
residential living needs (Snyder, Wallace, Moe, & Liberman, 1994).
Addressing these stressors by use of supportive therapy in the form of
individual, family, and group sessions has been found to reduce relapse
rates and improve social functioning (Hogarty, Goldberg, & Schooler, 1974).
Social skills training was found in one study to be more effective than
supportive therapy (Marder et al., 1996). Supportive group therapy with
varied emphasis on medication education, setting realistic goals,
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
encouraging coping efforts, and socialization were found to be superior to
other treatments (Malm, 1982).
Biopsychosocial Model and Substance
Abuse Problems
Theories linking substance abuse with antisocial and violent behavior
generally follow along one of three hypotheses. First, the Impairment
Hypothesis states that alcohol causes changes or impairment in the user
that make aggression more likely and/or more attractive (Bushman, 1997).
In a meta-analysis review of alcohol and aggression studies, Bushman
found evidence for the veracity of this model. He called it “Indirect Cause”
(p. 234). Aggression was more likely in certain social contexts.
Second, the Expectancy Hypothesis states that people who drink are
more aggressive because they expect to be or think they can get away with
the behavior and not be blamed due to the alcohol in their systems.
Bushman’s (1997) meta-analytic review did not support this hypothesis.
Third, the Spurious Hypothesis states that there is another element
not accounted for that co-occurs with the substance abuse and the anti
social behavior and that could explain the antisocial behavior instead of
substance abuse. In their review of violence and substance abuse litera
ture, Parker and Auerhahn (1998) concluded that there was no causal link
between alcohol and aggression. Cocaine and alcohol combined have
been found by other researchers to fit the Impairment or Indirect Cause
Hypothesis as the link between aggression and drug use (Denison,
Paredes, & Booth, 1997).
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In summary, the above evidence appears to be consistent with a
biopsychosocial model of criminality and violence in that it supports multiple
causes or multiple factors to be involved when substance abuse and
criminality co-occur.
Biopsychosocial Model and the
Presence of Psychopathy
The origin of psychopathy conceptualized by the biopsychosocial
model includes more than measures of past criminality. There is consistent
biological evidence that criminal behavior runs in families (Rutter & Madge,
1976). Twin studies and genetic studies show significant criminal herita-
bility (Cloninger, Reich, & Guze, 1978; Raine, 1993; Rowe, 2001; Rutter,
1997). Recent brain imaging studies have shown brain activity differences
elicited by emotional stimuli in diagnosed psychopathic versus nonpsycho-
pathic individuals (Andreason & Black, 2001).
Psychosocial evidence for the origin of psychopathy also comes from
twin studies that found stronger evidence for heritability of criminality than
have adoption studies (Raine, 1993). The stronger evidence of antisocial
behavior from twin studies indicated an interaction effect with the social
environment that was not present in adoption studies. It seems evident that
genetic processes interact with an environmental switch that turns on or off,
depending on the family environmental influences. A striking example of
this is from a Swedish male adoption study (N = 862) by Cloninger,
Sigvardsson, Bohman, and von Knorring (1982). The adoptees were
divided into four groups based on the presence or absence of (a) a con
genital predisposition with both biological parents being criminals and
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(b) a postnatal predisposition marked by how the children were raised.
There was a 40% rate of criminality when both biological and environmental
influences were present. There was only an 18.8% criminality rate when
only one influence, congenital only or postnatal only, was present.
Another biological factor for psychopathy could be involved when
comorbid attention-deficit/hyperactivity disorder (ADHD). Long-term follow-
up studies have shown that 30% of ADHD children developed criminality
(Weiss & Hechtman, 1993; West & Farrington, 1973).
Polloch et al. (1990) found that parental alcoholism, together with
physical abuse of a child, predicted that the child would develop adult
antisocial behavior. Psychopathy has been shown to develop in children
whose parents fail to discipline and supervise their activities (Robins, 1966;
West& Farrington, 1973).
Epidemiological research (Robins & Regier, 1991) has shown that
the prevalence of Antisocial Personality Disorder (APD; a behavioral marker
for psychopathy) was strongly correlated with age, gender, and socioeco
nomic status. Younger age, being male, and in a lower socioeconomic
class predicted antisocial behavior.
Another societal factor was reported in the ECA study (Robins &
Regier, 1991), which showed that in the United States the lifetime
prevalence of APD had nearly doubled in the past 15 years to 2.4%. These
changes could be accounted for only by the changes in the social environ
ment. By contrast, studies done in Taiwan in rural and urban areas showed
very low rates of APD in the range from 0.03% to 0.14% (Compton et al.,
1991; Hwu, Yeh, & Change, 1989).
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In his essay on psychopathy and the biopsychosocial model Paris
(1998) commented on the above Taiwan study: “ The low rates in Taiwan
are most probably due to the high levels of cohesion in the traditional
Chinese families” (p. 237). In the Robins (1966) study there was also a
very low APD rate in Jewish families, which she also attributed to strong
family structure.
In summary, there have been many theories of psychopathic human
behavior, substance abuse, psychotic mental illness, and criminal behavior.
To date, no overarching theory encompasses all of these interacting vari
ables. Instead, many “minitheories” address one or a few of the variables.
The biopsychosocial model theorizes that psychopathy develops
only in those individuals already vulnerable by genetic and environmental
influences and with personality traits of high impulsivity and high behavioral
activation (Raine, 2002). A biopsychosocial model of psychosis involves
biogenetic and psychosocial history assessment of schizophrenic patients.
Neuroleptic medications are prescribed to reduce symptoms and combined
with cognitive behavioral individual and group therapy focused on psycho
social skill development.
Theories of crime have relevance to the reduction of criminal
behavior. Theories of mental illness have relevance when they reduce the
patient’s pain and discomfort. Theories of interpersonal conflict have
relevance when problems between people are resolved and/or acceptable
alternatives are agreed upon.
The biopsychosocial model is appropriate for treating criminal
behavior among schizophrenics in much the same manner as described
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
above. The “punishing” effect of court constraints, together with the
rewards given for prosocial behaviors and increasing self-care, are often
convincing and effective means of shaping new, more adaptive behaviors
with the mentally ill offenders.
The Psychopathy Construct
A brief history of the construct and present use of the psychopathy
construct is presented, especially as it relates to the Hare PCL-R and the
PCL:SV used in this study and to the three mini-theories of psychopathy.
Then follow reviews of theories of drug abuse and violence, and active
psychosis symptoms as predictors of violent reoffense.
The construct “psychopathy,” as predominately used in contempor
ary forensic psychological literature, derives primarily from the work of
Robert D. Hare and his colleagues. Hare’s writings and descriptions
(especially the now well-known assessment tool, the Hare PCL-R) employ
many of the descriptors offered by Cleckley (1941,1982) in The Mask of
Sanity. Doren (1987) described Hare’s theory as a biologically based re
sponse perseveration. The three facets of Hare’s theory of psychopathy
are “(a) psychopaths have lesions in the limbic area of their brains, (b) such
lesions cause the loss of inhibitory mechanisms, and (c) such a loss brings
about the perseveration of the situationally dominant behavior” (p. 66).
Doren (1987) concluded that there is some positive supporting evi
dence in animal studies for Hare’s theory, assuming, of course, that human
brains and behaviors follow the same pattern and that all psychopaths have
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
limbic brain lesions. These assumptions have not been established in
human research.
The “ constitutional” or biologically based theory that Hare proposed
has both critics and supporters. Today, more of the forensic literature on
psychopathy focuses on the practical uses of the PCL-R and PCLSV tools
than on the theories of psychopathic personality. Nevertheless, it is
interesting to note that many early theorists expressed differing views when
explaining antisocial behavior. Some believed “ deviant” or antisocial
behavior to be based on constitutional and biological or genetically pre
determined factors. Others argued for more environmental or social learn
ing causes of deviant and antisocial acts. Pinel, Otto, Prichard, Koch,
Leary, and others have described such divergent opinions in the past 2
centuries (Millon, Simonsen, & Birket-Smith, 1998).
The working definition of a psychopath employed in this study draws
from characteristics similar to Cleckley’s definition (1982). He gave the
following list of 16 characteristics: (a) superficial charm and good “intelli
gence”; (b) absence of delusions and other signs of irrational thinking;
(c) absence of “nervousness” or psychoneurotic manifestations; (d) unrelia
bility; (e) untruthfulness and insincerity; (f) lack of remorse and shame;
(g) inadequately motivated antisocial behavior; (h) poor judgment and
failure to learn by experience; (i) pathologic egocentricity and incapacity for
love; (j) general poverty in major affective reactions; (k) specific loss of
insight; (I) unresponsiveness in general interpersonal relations; (m) fantastic
and uninviting behavior, with drink and sometimes without; (n) suicide,
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
rarely carried out; (o) sex life impersonal, trivial, and poorly integrated; and
(p) failure to follow any life plan.
Cleckley devised his list from observations of individuals rather than
from a rigorous objective study. His observations have shown remarkable
timelessness and functionality. Many of Cleckley’s listed items have found
a place in the now well-validated Hare PCL-R/PCLSV and therefore are
relevant to current forensic research.
Psychopathy History
Psychopathy literature has grown prolifically during the past decade;
this is especially true for studies that utilize the PCL-R. A brief review is
needed of the history and development of the construct psychopathy and
how it is used in this study, particularly, how it is assessed by the PCL-
R/PCL:SV. For a thorough reading of this history, the reader is referred to
Millon et al. (1998).
For the past 2 centuries, the topic of psychopathy has been
described by clinical characteristics and by its origins of behavior patterns.
Philippe Pinel (1801/1962) used the term la folie raisonnante to describe
impulsive and self-damaging acts by a person with unimpaired reasoning
abilities. Pinel also described these cases by the name manie sans delire
(insanity without delirium; Millon et al., 1998, p. 4).
In 1824 a Danish physician, Carl Otto, examined hundreds of con
victs in the Copenhagen prison. He wrote several books and medical
journals. In his view, “ criminal behavior was not due to the activity of any
single organ, but rather to a dysfunctional interplay of several organs, which
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
again was related to each individuals peculiar constitution and develop
ment” (Millon et al., 1998, p. 5).
The British alienist J. C. Prichard (1835) may have been the first to
use the concept “moral sanity.” He took Pinel’s idea of manie sans delire
and added the view that such behaviors showed a moral defect in character
and were deserving of societal condemnation. Due to Prichard’s work,
there was a split among British legal scholars. The two views of insanity
were (a) insanity due to defective reasoning ability, and (b) Prichard’s idea
of defects in “natural affections.” Prichard’s view was encompassing of
nearly all mental conditions and so does not, by his use and writing, appear
to be anything like our current clinical use of psychopathy or antisocial
personality disorder.
J. L. Koch (1891) was a prominent German psychiatrist who was
among those who turned from the pejorative English views of Prichard to
one of observational research. He used the term psychopathic inferiority.
By this he meant all congenital and acquired mental irregularities “ that influ
ence a man in his personal life and cause him, even in the most favorable
cases, to seem not fully in possession of normal mental capacity” (Koch,
1891, p. 67, as translated by Millon et al., 1998). He stated: “ They always
remain psychopathic, in that they are caused by organic states and
changes which are beyond the limits of physiological normality. They stem
from a congenital or acquired inferiority of brain constitution” (p. 54, as
translated by Millon et al.). Emil Kraepelin wrote about psychopathy in his
many editions of Psychiatrie: Ein Lehrbuch (Psychiatry: A Textbook). In his
last edition he described the psychopath as having behaviors that the DSM
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
has ascribed to juvenile conduct disorder. But his contemporaries and
those who followed disagreed with him. The discussion, often heated, has
never ended among proponents of various theories and ideas of psycho
pathy.
Hans J. Eysenck (1964) proposed a theory of psychopathy based on
a three-dimensional model of personality. In his model he specified:
There are certain major personality variables, independent of each
other; that these are in great measure genetically determined; and in
conjunction they can be used to allocate a given person (whether
psychiatrically well or ill) to a particular point in this multidimensional
space. (Eysenck & Eysenck, 1978, p. 198)
His first dimension was called extroversion-introversion (E). An
idealized extrovert is a person who loves parties and excitement, acts on
impulse, is carefree, and tends to be aggressive or lose his temper easily
(p. 50). On the opposite extreme is the idealized introvert, who is intro
spective, not impulsive, and emotionally controlled and reliable. The
second dimension was called neuroticism-stability (N). A highly neurotic
person overreacts to stimuli. If the reaction is overt, this loads on extrovert;
if covert, this loads on introvert. The third factor was that of psychoticism
(P). While this third factor helped to make his theory more comprehensive,
yet it lost is pointedness. He also called this third dimension “ tough
mindedness,” but he preferred the term psychoticism. He viewed psycho
pathy as “half-way stage to psychosis . . . a dimension of personality which
leads from outright psychosis through psychopathy to normality” (Eysenck,
1977, p. 57). Based on his investigations he characterized persons high on
psychoticism as
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(1)solitary, not caring for other people; (2) troublesome, not fitting in,;
(3) cruel, inhumane; (4) lack of feeling, insensitive; (5) lacking in
empathy; (6) sensation-seeking, avid for strong sensory stimuli; (7)
hostile to others, aggressive, (8) liking for odd and unusual things;
(9) disregard for dangers, foolhardy; (10) likes to make fools of other
people, and to upset them. (p. 58)
Eysenck’s theory has some empirical merit, but critics have ques
tioned the validity of some of the scales of his personality questionnaires,
which he authored. One of his statements does seem to have been sup
ported in laboratory work with human subjects. That is, there appears to be
strong support for psychopaths to have low cortical arousal. His conclusion
was that psychopaths have a low ability to be conditioned instrumentally or
through classical conditioning.
While this brief discussion does not do justice to Eysenck’s theory,
some of the findings have made their way into the ideas of others—
particularly Robert Hare, and specifically as Dr. Hare’s PCL-R is today the
most widely used and accepted tool to assess psychopathy. Eysenck said
that psychopaths inherited an extroversion temperament that predisposed
them to develop antisocial behaviors. There appear to be many unverified
elements to his theory. Nevertheless, his theory assumes a constitutional
disposition. Other theorists have posed vicarious learning and reinforce
ment as means for the sociopath or psychopath to develop aggressive
behavior. Eysenck’s theory has questionable utility at this time due to its
omissions and contraindications (Doren, 1996).
Bandura and Walters (1963) followed a social learning model and
proposed that the parent-child relationship is at the core of these deviant
social behaviors and attitudes.
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Epidemiology of Psychopathy
In recent decades, epidemiological research has utilized very reliable
Danish government records as sources for the study of psychopathy and
criminality. One such researcher, Fini Schulsinger (1972), from a sample of
5,483 identified 57 adoptees hospitalized with the diagnosis psychopathy
and matched them with 57 control nonpsychopathic adoptees. It was found
that twice as many biological relatives as adoptive relatives suffered from
psychopathy. There were no identified environmental factors.
Several studies conducted by Mednick and colleagues (Hodgins,
Mednick, Brennan, Schulsinger, & Engberg, 1996) examined the genetics
of criminal behavior and looked for an association between crime and
psychopathy. One study looked at a sample of 4,065 adopted males. The
highest group of adopted males with one criminal conviction was found in
the group with both biological and adoptive criminal parents. The next high
est group with one conviction was in the group in which only the biological
parents were criminal. In sum, these twin studies yielded evidence of a
genetic or biological link to criminal behavior and psychopathy.
Cleckley’s (1941,1982) observations and list of behaviors have for
many years served as the recognizable syndrome of psychopathic
behavior. Many items from Cleckley’s list are recognizable in the current
20-item Hare PCL-R and the shorter, 12-item screening version (PCL:SV).
The screening version was developed later in connection with the
MacArthur Foundation’s ECA research (Robins & Regier, 1991).
Hare developed his theory of psychopathy from studies of rats’
behavior. He observed that limbic lesions in the rats caused them not to
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
learn from punishing consequences. He hypothesized that it was limbic
lesions in human brains that caused them to fail to change behavior when
punishing mistakes were made. This biologically based limbic lesion theory
has been challenged. Recent brain scan research (Andreason & Black,
2001) has shown some evidence of a biological basis for psychopathy.
Biological research continues to look for the genetic link to antisocial and
psychopathic behavior
Schizophrenia/Psychosis and Violence
Several research investigations have shown evidence of some link
between mental illness and violent crime. The present author’s 12 years of
clinical observation with clients on conditional release from state hospitals
also supports the mental illness variable as a predictor of future reoffense
as well. Monahan (1992) suggested that florid psychotic symptoms are
more highly correlated with violent behavior than is the simple diagnosis of
a psychotic disorder.
In their quantitative review, Douglas and Hart (1996) found a
stronger effect during analysis for specific psychotic symptoms versus
gross categories of mental illness. Link and Stueve (1994) and Swanson et
al. (1996) found that psychotic symptoms, which both override one’s sense
of safety or well-being (so-called “ threat/control-override” symptoms), were
more strongly related to violence than psychotic symptoms, which do not
possess these qualities. De Pauw and Szulecka (1998) claimed some evi
dence that patients “ with well-developed delusions are more likely to
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
commit violent crimes against persons than those with chronic undifferen
tiated psychoses” (p. 91).
Many studies have observed a relationship between psychiatric
symptoms, as measured by the BPRS (Overall & Klett, 1962) and violence
(Douglas & Hart). Other clinical conditions that should be accounted for
include such states as sadistic fantasies (MacCulloch et al., 1983), suicid-
ality (American Psychiatric Association, 1974; Brent et al., 1994; Hillbrand,
1995), and paranoia, self-aggrandizement, and pathological jealousy.
Homicidal ideation clearly must be noted and investigated.
Linking active psychosis, specifically delusions and hallucinations, to
criminal behavior could be found to be very similar to linking any mental
illness to criminal behavior. To look at delusions and hallucinations more
narrowly defines behavioral symptoms.
One relevant theory of criminal behavior is Social Learning Theory
(SLT; Akers, 1985). SLT has long been considered by some criminologists
as one of the best operational theories for the understanding of recidivism.
SLT is relevant to treatment in that the theory states that all behavior is
learned and can therefore be unlearned (Dalton, 2000). Akers (1998)
expanded his theory and called it Social Learning Social Structure (SLSS;
see earlier explanation of Akers’s SLSS theory in the theory section).
Link and Stueve (1994) proposed a model to explain a link between
mental illness and violence. In this model they stated that the behavior of
the psychotic patient is rational, or is predictable, if one first knows what is
the perceived reality of the patient. For example, a patient would be
expected to defend or fight for his or her life if a mortal threat were believed
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
to be present. When the threat is perceived but not real, it is labeled a
paranoid delusion. An interesting note is that this model emerged from a
study and analysis undertaken to show that no significant link exists
between mental illness and criminal behavior. The authors stated in this
model that a “rationality” or understandable explanation of a violent act
occurs with some psychotic and irrational symptoms.
The principle of rationality-within-irrationality posits that once one
suspends concern about the irrationality of psychotic symptoms and
accepts that they are experienced as real, violence unfolds in a
rational fashion. By rational we do not mean reasonable or justified
but rather understandable. (Link & Stueve, 1994, p. 143)
For example, when a person fears personal harm or feels threatened by
others due to delusions or hallucinations, violence is more likely to follow.
Link and Stueve (1994) investigated this model using data collected
by Dohrenwend and colleagues, who used the PERI (Dohrenwend et al.,
1986; Shrout et al., 1988). Link and Stueve developed a TCO scale using 3
of the 13 items of the PERI’s psychotic symptom scale. The other 10 items
were labeled “ other psychotic symptoms."
These scales were used to analyze the results of three questions
posed to three separate population samples. Each group was matched
between psychiatric patients and nonpatient community participants. The
first group answered the question about “hitting” in the previous month or
year, the second group answered the question about “ fighting” in the
previous 5 years, and the third group answered the question about “using
a weapon” in the previous 5 years.
Due to the way in which the above research information was
collected, the three data sets were analyzed separately. The first group,
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
“hitting,” showed that 12.3% of patients (n = 385) reported hitting in the
previous year, while 5.2% of the nonpatients (n = 365) reported hitting in
the same time period. Of the second group, “ fighters,” 25.7% of the
patients (n = 191) and 15.1% of the nonpatients (n = 185) reported fighting
in the previous 5 years. In the third group, “used a weapon,” 9.7% of the
patients (n = 195) and 2.7% of the nonpatients (n = 186) reported using a
weapon in the previous 5 years. Utilizing further regression analysis, the
TCO scale remained a significant predictor of violence. This was true even
when the other psychotic symptoms were held constant. The same pattern
of significance for TCO symptoms remained after other possible
determinants of behavior were included in the analysis.
Dual Diagnosis and Violence
The relationship between substance abuse and violence in the
general population has long been debated. Substance abuse has been
shown to be a significant factor in reoffense among seriously mentally ill
offenders in the large ECA study (Swanson, 1994). Among the mentally ill,
there was a 7% absolute risk of violence in the course of a year. Among
single-diagnosis substance abusers, there was a self-reported prevalence
of violence of 55.2%. Among those with mental illness and substance
abuse (dual diagnosis), the rate was 63.9%.
Substance abuse problems proved a significant predictor of violence
in the study by Harris et al. (1993). In their recent reviews, Klassen and
O’Connor (1994) regarded substance abuse as an important link to violent
behavior. Other research supports the link between substance use and
61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
violence in various sample populations (Bartels et al., 1991; Blomhoff et al.,
1990; Hodgins, 1990; Hoffman & Beck, 1985; Steadman et al., 1998;
Taylor, 1985).
A growing body of research has established an association of
criminal or violent offense with dual diagnosis of major mental illness
(thought disorders) and substance abuse disorders. Causal links between
the dually diagnosed and criminal behavior have remained elusive,
although frequently hypothesized. The ECA study (Robins & Regier, 1991)
has shown a significant correlation between the dually diagnosed (mental
illness and substance abuse disorders) and criminal violence and non
violent offenses. To date, a thorough and convincing theoretical basis for
such an association or causal link between the dually diagnosed and
criminal violence has not been established. Several theories have been
proposed in the literature. Two theories are briefly described below.
The first theory of a link between substance abuse and violence is
Fagan’s (Fagan, 1993; Fagan & Chin, 1990) “ contextual approach.” Fagan
attempted to devise a comprehensive model for the relationship between
violence and substance abuse. He reviewed theoretical arguments from
psychological and psychiatric research, psychopharmacological
approaches, social and cultural viewpoints, and biological and physiological
research. The following is a brief summary of his synthesized model.
1. Intoxication impacts thinking ability and functions, which vary by
substance.
2. The context of the behavior influences how people function:
(a) social and cultural meanings, (b) understanding about the impact of
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
intoxication on judgment, (c) ability to perceive social cues, and (d) ability
to focus on long- and short-term outcomes of situations involving substance
presence, including a possible violent outcome.
3. The setting where the interaction takes place influences violent
outcome.
4. The presence or absence of social controls influences violent
outcome.
5. The intoxicated person’s limited response set to social situations
influences violent outcome.
For youth violence, Fagan and Wilkinson (1998) proposed “ a
functional, purposive behavior that serves definable goals within specific
social contexts” (p. 2). Attainment of status is the goal and benefit of youth
violence.
The second theory of a link between substance abuse and violence
was proposed by Goldstein (1989,1995), naming it the “ tri-partite model.”
The three parts are (a) pharmacological— behavior directly influenced by
the chemicals in the body; (b) economic—crime and violence resulting from
the struggle to get money to buy the drugs that a person is craving; and
(c) systems—for example, organized crime operants who commit crime by
collecting money from those who bought the illegal drugs.
Summary of Hypothesized Predictor Literature
More and more empirical studies confirm the relationship between
psychopathy and increased criminal and violent offenses. Separate and
different populations confirm this finding. The meta-analytic reviews cited
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
lend collective and strong support for psychopathy as a main predictor of
reoffense and violent behavior. Among sex offenders, measures of psycho
pathy, combined with a measure of deviant sexual interest, predict sexual
offense. Recent studies using the 20-item PCL-R with schizophrenic
offenders have shown psychopathy to predict violent recidivism (Nolan et
al., 1999; Tengstrom et al., 2000).
Active psychosis as a predictor of reoffense has had mixed reviews
in the literature. A growing majority of studies have shown an increased
risk for reoffense and violence when command hallucinations are currently
present and then certain delusions are perceived as threatening (Swanson,
1994).
The dual-diagnosis studies show that mental illness, combined with
substance abuse, increases the risk for reoffense and violence more than
9-fold (Raesaenen et al., 1998).
In light of the literature reviewed, four hypotheses were tested.
1. Violent and nonviolent reoffenders will have higher psychopathy
then nonreoffenders.
2. Violent and nonviolent reoffenders will show signs of active
psychosis more than nonreoffenders.
3. Violent and nonviolent reoffenders will have more substance
abuse problems than nonreoffenders.
4. The 15 items on the Historical and Clinical scales of the HCR-20
will more accurately predict violent and nonviolent reoffense in this
population than the PCLSV.
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 3
METHOD
This chapter describes the methods employed in this study.
Included are (a) descriptions of the participants (patients), (b) data
collection procedures and the measures utilized in data collection, and
(c) definitions of terms used in the study. The participant section includes
selected demographics of patients, crime history, and psychiatric diagnostic
descriptions. The data collection and measurements section describes the
measures used and the rating system employed. This section includes
information on raters (research assistants) and their professional back
grounds, how the raters were trained, the process of coding each patient
file, establishing interrater reliability, the setting in which the files were
reviewed, and the types or sections of file information from which the coding
was obtained. In the final section terms are defined operationally for
application to this study.
Participants
This study examined archival data of 53 Schizophrenic or Schizo
affective male inpatient offenders treated by the California DMH. All patient
data were from existing file records at Patton State Hospital at the time of
release into outpatient treatment and from the subsequent outcome
measures during the court-mandated outpatient treatment in CONREP. All
participants entered subsequent CONREP treatment as mandated by the
court. The uniform outpatient treatment given to all CONREP patients
includes the following “ core services”: (a) weekly individual and group
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
therapy, (b) weekly random drug screens, (c) monthly psychiatric medica
tion adjustments, (d) monthly home visits, and (e) bimonthly collateral con
tacts. The studied offenders are from the legal category NGRI described in
PC1026. NGRI means that the court declared that the patient was not able
to know right from wrong because of mental illness at the time that the
crime was committed. California’s NGRI patients historically average 4.5
years of inpatient treatment prior to obtaining a COT order by the court and
admission into a CONREP treatment program (Wiederanders & Choate,
1994). The average length of the participant patients’ inpatient treatment
prior to release to CONREP in this study was 3.7 years.
The goal of CONREP treatment is to maintain psychiatric stability
and keep active psychosis in remission, especially to eliminate hallucina
tions, delusions, and paranoia. While rendering these services, CONREP
supervises and monitors each patient for compliance with the court-ordered
“ Terms and Conditions” of outpatient treatment. These Terms and Condi
tions are roughly analogous to law enforcement’s use of parole rules in
community enforcement and management of parolees. This monitoring or
supervision is the forensic focus of treatment, and the purpose is the
protection of the community from any further violence or unlawful behavior.
Any violation of the court-ordered Terms and Conditions of outpatient
treatment results in a rehospitalization and court hearing regarding the
violation. This revocation hearing to end outpatient status is roughly
comparable to parole violation hearing.
The patient files for this study included the entire available patient file
population that met the selection criteria by diagnosis, legal class, and age
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
at entry to outpatient treatment from Patton State Hospital. More informa
tion is given below under File Selection.
The custodian of the DMH’s CONREP records generated and
identified the patient records reviewed in this study. The records consisted
of Patton State Hospital closed chart records, rearrest records, and
rehospitalization records.
Population Parameters
Several parameters were chosen with the goal of making the out
come of this study more meaningful and useful. The parameters are PC
1026 legal status (NGRI), diagnosis, age, exposure time to risk, and
gender.
Legal Status
All participants were of the legal class NGRI. This determination is
made by the court and states that the participant did not know right from
wrong when committing the crime because of mental illness.
Diagnosis
In order to secure an appropriately homogeneous diagnostic group
for this study, only patients with one of the diagnoses Schizophrenia or
Schizoaffective were included. This eliminated those patients with drug-
induced psychotic disorders and those who may have other psychotic
features but did not meet the formal criteria of one of the schizophrenia-
related diagnoses.
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Age
It has been long established that the onset of schizophrenia most
often begins from late adolescence to adulthood. In the Comprehensive
Textbook of Psychiatry Wyatt, Kirch, and Egan (1995) described the onset
of schizophrenia: “Identifiable symptoms of schizophrenia are rare before
puberty. The peak age at onset for males is between 18 and 25 years, and
for females between 26 and 45. Although schizophrenia can begin after
age 45, late onset is uncommon” (p. 927). Since schizophrenia has a
typical onset in late teens or early adult years, the lower age limit for this
study was set to begin at age 25. Males are thought to have developed
schizophrenia by this age and females as well. As noted above, some
professionals believe that schizophrenia may develop later in females than
in males. It is hypothesized that there is more scrutiny and certainty of the
schizophrenia diagnosis by setting age 25 as the lower age limit. The final
group of participants included two males age 23, as available participants
were disappearing by hospital transfer and because these individuals had
over 2 years of “ exposure to risk” in the outpatient program.
The upper age limit cut-off for time to enter CONREP was 38 years
for this study. This age was used due to the extensive literature stating that
criminal behavior tends to diminish with age. Some researchers have found
evidence that mental patients under age 40 were disproportionately violent
in the community (Klassen & O’Connor, 1994). Utilizing age 38 as the
upper age limit allows for the peak age of risk for criminality to continue
after a patient leaves the hospital and enters CONREP by age 38.
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Exposure to Risk
The researcher set no limit of time for exposure (or risk) for
reoffense. In other words, each patient who met the criteria of legal status,
diagnosis, and age was included in this study upon his release from Patton
State Hospital into CONREP COT treatment. The average time that a
patient remained in the mandated outpatient treatment setting was about 3
years.
Gender Issues
The database used for this study identified only 3 female schizo
phrenic patients who met the parameters established: age, diagnosis, and
legal status. There has been a gradual increase in the number of female
patients in treatment. For example, in 1990 approximately 1 in 13 patients
in treatment were female. In 2000 approximately 1 in 9 patients were
female. The literature is increasingly reporting studies relating to female
violence and reoffense. Several researchers have reported little or no
difference between male and female patients who are violent (Hiday,
Swartz, Swanson, Borom, & Wagner, 1998; Rabinowitz & Mark, 1999;
Tardiff, Marzuk, Leon, Portera, & Weiner, 1997).
There were 53 male patients who met the parameters and were
included in the study. The 3 female patients were excluded from the study
due to being too few in number.
File Selection
Originally, 116 file numbers that met the requirements were identi
fied. However, after the coding began, it was learned that some files had
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
been destroyed because they had been inactive for 84 months. Others
were not available because the patient and chart had been transferred to
another hospital. One patient had died with only 1 month of exposure and
was eliminated from the selection. Three files were found to contain a
misdiagnosis in the database, and three were found to be over the age of
40 years when released into outpatient treatment. Three female patients,
for the reasons stated above, and 5 male patients, because the reason for
rehospitalization was not stated, were excluded from the study, resulting in
a total of 53 patient files that were used for the data set analysis. All of the
files reviewed were from a single state hospital in southern California, which
accepts forensic psychiatric patients from throughout the state.
Data Collection and Measurements
The data for analysis were collected and coded as described above,
and code sheets were completed for each of the 61 patient charts reviewed.
The two instruments used were the PCL:SV and the HCR-20. The five
items of Risk Management were excluded from the HCR-20, making this
what is hereinafter termed the “HC-15.”
Instrumentation
There is wide acceptance of the validity and reliability of the PCL-
R/PCLSV among forensic evaluators, and this checklist is fast becoming
the hallmark of assessing or diagnosing “psychopathy.” Some of the major
support for this acceptance is noted herein. The HCR-20 has grown in
usefulness and acceptance among forensic evaluators since it was first
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
developed in 1995. It has recently received attention for its utility in
treatment planning with forensic psychiatric clients.
The PCL:SV and the HCR-20 have shown promise and accuracy in
violence prediction in the 70+% range. There is still much room for
improvement in the accuracy figure.
P C L S V
The PCLSV is a 12-item version of the longer, 20-item Hare PCL-R
(Hare, 1991,1998). The PCL-R is a rating scale designed to measure
psychopathic personality traits in forensic populations. Each of its 20 items
reflects a different characteristic of psychopathy or trait. The items are
defined in detail in the PCL-R manual and are rated on a 3-point scale (0 =
item doesn’t apply, 1 = item applies somewhat, 2 = item definitely applies).
A sum of the items yields a total score between 0 and 40. The score
indicates the degree to which an individual fits the profile of the prototypical
psychopath. A score of 30 has been used as the cutoff to diagnose
psychopathy.
The PCL-R items also can be summed to yield scores on two
moderately correlated factors (Cooke, 1995; Hare et al., 1990; Harpur,
Hakstian, & Hare, 1988; Harpur, Hare, & Hakstian, 1989). The first reflects
the affective and interpersonal features of psychopathy and has been
labeled the Selfish, Callous, and Remorseless Use of Others. The second
reflects the social deviance features of psychopathy and has been labeled
Chronically Unstable and Antisocial Lifestyle (Hare et al., 1990).
71
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A large and impressive body of research has evolved for the PCL-R
as it was originally developed for use in experimental psychopathology (see
Cooke, Forth, & Hare, 1998 for a review). More recently, however, the
PCL-R has received much attention for its predictive validity with respect to
criminal behavior, particularly violent crime (Hart & Hare, 1996; Salekin et
al., 1996). Studies have associated PCL-R results and conditional release
failure (Hart et al., 1988; Hemphill, Hare, et al., 1998; Serin, Peters, &
Barbaree, 1990), violent recidivism (Harris et al., 1993; Quinsey et al.,
1995; Serin, 1996; Serin & Amos, 1995), and poor treatment response
(Ogloff, Wong, & Greenwood, 1990; Rice, Harris, & Cormier, 1992).
The PCL-R has received favorable comments on its psychometric
properties and its criterion construct-related validities (Darke, Kaye, Finlay-
Jones, & Hall, 1998; Hare, 1991; Hobson & Shine, 1998). However,
administration requires access to detailed interview and case history
information, including a criminal record. This becomes costly, especially in
terms of time and effort. Consideration of these issues led to the develop
ment of the PCL:SV (Hart, Cox, & Hare, 1995).
The PCLSV concisely collapses the 20 items of the PCL-R into a
12-item rating scale (Hart et al., 1995). Six items reflect Factor 1 (affective
and personality features) of the PCL-R, and six items reflect Factor 2 (social
deviance behavior) of the PCL-R. The item descriptions in the PCL:SV
manual are very brief, relative to those for the PCL-R, and require less
detailed information to score. The same 3-point rating scale is used for the
PCL.SV as for the PCL-R. A cut-off score of 18 is suggested to diagnose
72
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
psychopathy. Research presented in the manual for the PCLSV suggests
that file-only data tend to underdiagnose psychopathy.
The first six items of the PCLSV are Superficial, Grandiose, Deceit
ful, Lacks Remorse, Lacks Empathy, and Doesn’t Accept Responsibility.
These six items are called factor one and reflect the “ severity of the inter
personal and affective symptoms of psychopathy” (Hart et al., 1995, p.22)
The last six items are Impulsive, Poor Behavior Controls, Lacks Goals,
Irresponsible, Adolescent Antisocial Behavior, and Adult Antisocial
Behavior.
Two examples from the 12 items are as follows: Item 1, Superficial
ity, is the first example. A score of 2 for this item means that this symptom
is definitely present and uses evidence from two or more separate sources
that two or more of the following characteristics are true for the person.
These are such things as (a) the presentation is shallow or superficial,
(b) the person portrays himself in a positive light and/or engages in
impression management, (c) several conflicting impressions are in the case
record, or (d) the person fails to answer the question but fills in the time with
his/her own agenda. Item 11, Adolescent Antisocial Behavior, is the
second example. This item refers to serious contact with the criminal
justice system before the age of 18. The score is 0 if no arrests or self-
reported charges are present. The score is 1 for minor offenses such as
possession of drugs, minor theft, vandalism, or minor driving violations.
The score is 2 for serious offenses such as violence, major driving
violations, or drug trafficking.
73
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The PCLSV is both a reliable and valid measure of psychopathy
(Cooke, Michie, Hart, & Hare, 1999; Hart et al., 1995). It has been found to
correlate highly (r> .80) with the PCL-R in samples administered on separ
ate occasions by independent raters (Hare, 1991; Hart et al., 1995). These
samples include civil psychiatric patients, prison populations, and criminal
psychiatric patients (Douglas et al., 1999; Hart et al., 1994; Hill, 1997).
HCR-20
The HCR-20 is a violence risk assessment measure developed by
Webster, Douglas, Eaves, and Hart (1997). This assessment device was
constructed to be applicable to a variety of populations, including civil and
forensic psychiatric patients and correctional offenders. Borom (1996)
reviewed the HCR-20 in the American Psychologist, commenting, “ The
promise of this instrument lies in its foundation on a conceptual model or
scheme for assessing dangerousness; its basis in the empirical literature;
[and] its operationally defined coding system. The field eagerly awaits new
data on this instrument” (p. 950). This instrument is available as a manual
(Webster, Eaves, Douglas, & Wintrup, 1995; Webster et al., 1997). The
HCR-20 contains three scales (H, C, and R) and 20 variables noted below.
The Historical (H) scale contains 10 items of a generally static
nature: (a) previous violence, (b) young age at first violent incident,
(c) relationship instability, (d) employment problems, (e) substance abuse,
(f) major mental illness, (g) psychopathy, (h) early maladjustment, (i) per
sonality disorder, and (j) prior supervision failure.
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The Clinical (C) scale has five items and refers to current mental,
emotional, and psychiatric status. These “mental status” risk markers are
considered dynamic and sometimes changing from one time to another:
(a) lack of insight, (b) negative attitudes, (c) active symptoms of major
mental illness, (d) impulsivity, and (e) unresponsive to treatment.
The Risk Management (R) scale has five items and is concerned
with forecasting the future social, living, and treatment circumstances of an
individual. It also attempts to anticipate the person’s reactions to social,
living and treatment conditions: (a) plans lack feasibility, (b) exposure to
destabilizers, (c) lack of personal support, (d) noncompliance with remedia
tion attempts, and (e) stress.
Scoring is similar to and borrowed from the scoring of the PCL-R
(Hare, 1991) and consists of the following scale: 0 = available information
contraindicates the presence of the item, 1 = available information possibly
suggests the presence or mild presence of the item, and 2 = available
information definitely indicates the presence of the item.
There has been some limited research using the HCR-20. Wintrup
(1996; see also Douglas, Webster, & Wintrup, 1996) coded the HCR-20
and the PCL-R on a sample of 80 male forensic psychiatric patients. These
patients were released from a maximum security forensic inpatient institute
and followed into the community. Correlations between both the HCR-20
and the PCL-R and violence averaged a little below r= .30. The HCR-20
was more strongly linked to hospital readmissions for forensic reasons (r >
38) and psychiatric hospitalizations (r= .45) than to follow-up violence per
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
se; the PCL-R was not as strongly related as the HCR-20 to these indexes
(r= .25 and .36, respectively). Interrater reliability was not reported.
In a study of 279 involuntarily committed psychiatric patients
(Douglas et al., 1999; see also Douglas, 1996), researchers investigated
the relationship of both the HCR-20 and the PCL:SV with future community
violence. Patients were followed for 2 years. Interrater reliability for the
HCR-20 total scale was .80, using an intraclass correlation coefficient. In
this study the HCR-20 scores were the more stable predictors of the rates
of various indexes of community violence. The H scale was most consist
ently and strongly related to outcome. Through hierarchical regression
analysis it was found that the HCR-20 added predictive ability beyond the
PCL.SV, but the PCL:SV did not add predictive ability to the HCR-20.
Raters and Rater Training
A trained team of four mental health professionals coded the two
instruments retrospectively from existing charts and records on the selected
individuals who qualified for inclusion. Two additional mental health pro
fessionals gathered the demographic information from each file record. The
team of four instrument raters consisted of one female licensed psycho
logist with over 15 years of experience, one female licensed marriage and
family counselor with over 20 years of experience, and one male and one
female psychology intern, each with over 3 years experience. The addi
tional two demographic information gatherers were one male marriage and
family therapist with 20 years of experience and one female licensed
psychologist with over 25 years of experience. All raters had experience
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
with treating mentally ill offenders like those in CONREP treatment and
were trained in the use of the two instruments, PCLSV and HCR-20. Dr.
Robert Hare and Dr. Adelle Forth also formally trained three members in
the use of the PCL-R during a 3-day intensive training seminar. This PCL
training served as the foundation for the scoring of the PCLSV, and the
same scoring system was used for the HCR-20 scoring. The manuals for
the PCL-R:SV and the HCR-20 were followed for the scoring of participant
records. The HCR-20 was shortened to 15 items by removing the five Risk
Management items. This was done because there was insufficient
information in the records to score the Risk Management items. These five
eliminated Risk Management items would typically be assessed after the
exact circumstances of the community placement are known. The hospital
records did not contain this type of information. Three items from the HC-
15 were used to detect the presence of the three independent variables of
this study: psychopathy, psychosis, and substance abuse.
After training was received from Dr. Hare and Dr. Forth, the research
team took additional scoring training. This training included coding 10
simulated chart files that were not included in the study, in order to establish
interrater reliability agreement and reliability (Tinsley & Weiss, 1975). Inter
rater agreement and reliability were established after the first three charts
were independently reviewed and coded by two of the evaluators. Then the
other four members of the team independently evaluated another set of
three charts. Two of the three charts were within the recommended range
of differences, which is a score difference of 2. The remaining 4 of these 10
charts were then evaluated by all of the evaluators independently.
77
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Interrater agreement and reliability were within the recommended limits of
the manuals for the two instruments. Interrater reliability was very high, R =
.98 for the PCLSV and R = .97 for the HC-15 (Tinsley & Weiss, 1975). The
recommended interrater agreement of a maximum difference of 2 in the
total score was present from the beginning after the initial interrater trials
were completed. Later in the chart review and coding, a check was made
to assure that interrater agreement and reliability were maintained.
The raters were given a list of charts to review, a few at a time due to
the limited space available at the on-site location. In each chart, specific
documents were carefully reviewed initially in order to locate the available
records yielding the most information for coding. These sections included
the initial intake records, social histories, psychological evaluations, criminal
records, and the last 6 months of daily behavior notes. Other sections of
particular interest were psychiatric evaluations, BPRS assessments,
criminal history reports, and hospital incident reports as well as assess
ments made by the visiting CONREP staff who had followed each patient
until he was released to the CONREP program.
Coding Criteria for Independent Variables
The independent variables psychopathy, psychosis, and substance
abuse were measured by three items with these names from the Historical
and Clinical scales of the HC-15. The criteria used for coding these items
are found in the HCR-20 manual (Webster et al., 1997). The criteria used
for coding psychopathy on the PCL:SV are found in the PCL:SV manual
(Hare, 1991). Number codes and criteria are summarized in Table 2.
78
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 2
Coding Criteria for the Independent Variables
Variable
Code
score Criteria for code score
Psychopathy 0 Nonpsychopathic. Score of under 13 on the
PCLSV
1 Possible/less serious psychopathy. Score of 13-
17 on the PCLSV
2 Definite/serious psychopathy. Score of 18-24 on
the PCLSV
Psychosis 0 No active symptoms of major mental illness in
the last 6 months before hospital release
1 Possible/less serious active symptoms of major
mental illness in the last 6 months before
hospital release
2 Definite/serious active symptoms of major
mental illness in the last 6 months before
hospital release
Substance Abuse 0 No history of substance use problems
1 Possible/less serious history of substance use
problems
2 Definite/serious history of substance use
problems3
Note. PCLSV = Psychopathy Checklist-Screening Version.
aEvidence for a score of 2 for this variable was determined if the record
contained a formal substance abuse diagnosis or reported functioning
impairment history at home, at work or in the community due to substance
use.
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The presence of psychopathy was measured by the coding of item
H7, Psychopathy, on the HC-15. The presence of active symptoms of
psychosis in major mental illness was measured by the coding of item C3,
Active Symptoms of Major Mental Illness, of the HC-15. (The raters did not
use suicidality or signs of depression as active symptoms for any of the
participants in this study. Only delusions, hallucinations, and paranoia were
coded when present in the records.) Evidence for the active psychosis
variable was taken from the records of the last 6 months of hospital treat
ment before release to CONREP outpatient care. Sources for evidence
included BPRS scores, daily treatment notes, discharge summaries, recent
medication changes due to the return of psychiatric symptoms, recent
formal treatment summaries, and recent formal reports. The presence of
substance abuse problems was measured by using the coding criteria of
item H5, Substance Abuse Problems, of the HC-15.
Coding Outcome Dependent Variables
After this “before” team of raters completed the chart reviews,
another independent investigator matched “after” or outcome information,
which included the current status of the 53 charts reviewed. The outcome
data came from hospital charts, Department of Justice arrest records, and
the DMH Conditional Release database. The person matching these
records of “before” and “after” information did not participate in the coding of
the files at the beginning of the study, to assure that there would be no bias
by wishful intent. The outcomes assessed were (a) success (no new crime,
not rehospitalized); (b) violent offense (any physical violence including
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
armed robbery, arson, assault, murder, sex assault, use of weapon to make
a terrorist threat); and (c) nonviolent offense (includes all other reasons for
revocation such as theft of property, forgery, treatment noncompliance,
drug abuse, and psychiatric decompensation).
The design of this study was a correlation investigation of coded
patient characteristics compared to behavior outcome over time. The
variables of interest were (a) psychopathy, (b) substance abuse history,
(c) presence of active psychosis upon release from the state hospital,
(d) prediction accuracy of the PCLSV regarding violent and nonviolent
reoffense, and (e) prediction accuracy of the 15 items (HC-15) of the HCR-
20 regarding violent and nonviolent reoffense. Statistical analysis included
the use of standard statistical analysis computer software. Correlation
tests, regression analysis, and receiver operating curve processes were
completed. The ROC analysis determines the confidence level and
specificity level at different cut-off points of the two instruments (PCLSV
and HCR-20/HC-15) used for prediction of violence.
Definition of Terms
Active psychosis is coded according to the presence or absence of
active psychotic-type symptoms such as hallucination, delusions, paranoia,
sadistic fantasies, self-aggrandizement, pathological jealousy, suicidality,
and homicidal ideation.
Nonviolent offense includes offenses not threatening or producing
personal injury, including property crimes, theft, forgery, vandalism, and
81
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
rehospitalization due to treatment noncompliance, AWOL, or psychiatric
decompensation.
Psychopathy is a personality and behavior diagnostic term used and
applied by the guidelines set out in the manual for the PCLSV. The cut-off
score for psychopathy is recommended as 18 (see PCLSV manual for
scoring and item analysis).
Substance abuse as a marker includes any history of abuse that
could be diagnosed using DSM-IV (American Psychiatric Association, 1994)
criteria.
Violent offense includes all crimes and acts against persons, includ
ing murder, rape, assault, robbery, arson, residential burglary, and any
crime using a weapon.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 4
RESULTS
The results of this study are presented first in descriptive statistics,
followed by results of the tests of the four hypotheses listed in chapter 1.
Descriptive Statistics of the Sample
Table 3 presents a summary of the demographic statistics of the
sample. The age range upon admission to CONREP was 23-38 years, with
a mean of 31.3 years. The ethnicity distribution was as follows: 64.2%
White, 15.1% African American, and 20.8% Hispanic. All participants were
male. Regarding marital status, 73.6% were single and never married,
3.8% were married, and 22.6% were divorced or separated.
The diagnoses distribution was as follows:
1. Axis I disorders: Schizophrenia = 38 and Schizoaffective = 15
(total 53);
2. Substance abuse diagnoses: alcohol abuse = 3, cocaine = 1,
marijuana = 1, amphetamine = 1, polydrug abuse = 12, and caffeine
intoxication = 5 (totals 3 abusing alcohol and 15 abusing all other drugs,
excluding caffeine); and
3. Axis II disorders: Antisocial Personality disorder = 16, Border
line Personality disorder = 3, Borderline Intellectual Functioning = 1,
Dependent Personality disorder = 1, Schizotypal Personality disorder = 1,
Personality Disorder NOS = 3 (total 21 Axis II diagnoses, 32 No Diagnosis
on Axis II).
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 3
Demographic and Diagnostic Characteristics of Sample (N
Characteristic and category n
Gender
Male 53
Female 0
Age at discharge (years)
Length of index hospitalization (months)
Time in CONREP until current
date or revoke date (months)
Ethnicity
Caucasian 34
Black 8
Hispanic 11
Highest education level (insufficient information)
Marital status
Married or common law 2
Never married 39
Divorced or separated 12
00
4^
53)
% Range Mean SD
100.0
0.0
23.0-38.0
3.2-137.2
31.3
44.7 33.9
64.2
15.1
20.8
2.0-84.0 33.7 23.1
3.8
73.6
22.6
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 3 (continued)
Characteristic and category
Criminal history (instant crime)
Violent
Murder/manslaughter
Physical assault
Robbery/burglary
Sexual offense
Arson
Kidnapping
Nonviolent
Theft of personal identity
Schizophrenia diagnosis (n = 53)
295.3x Paranoid Type
295.9 Undifferentiated
295.6 Residual
295.7 Schizoaffective
Substance abuse diagnosis
303.90 Alcohol Dependence
305.00 Alcohol Abuse
304.4 Amphetamine
Dependence
305.70 Amphetamine Abuse
305.90 Caffeine Intoxication
304.20 Cocaine Dependence
305.20 Marijuana Abuse
304.80 Polydrug Abuse
n % Range Mean SD
20 37.7
20 37.7
5 9.4
4 7.5
2 3.8
1 1.9
1 1.9
27 50.9
10 18.9
1 1.9
15 28.3
2 3.8
1 1.9
1 1.9
1 1.9
5 9.4
1 1.9
1 1.9
12 22.6
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 3 (continued)
00
CD
Characteristic and category n % Range Mean SD
Total with Drug Abuse diagnosis 24
Total excluding Caffeine Intoxication 19
Personality Disorder Diagnosis on Axis II (n = 21)
301.70 Antisocial Personality Disorder 16 30.2
301.83 Borderline Personality Disorder 2 3.8
301.2 Schizotypal Personality Disorder 1 1.9
301.9 Personality Disorder NOS 2 3.8
Total Personality Disorder 21
Other Diagnoses on Axis I and II
302.2 Pedophilia 1 1.9
302.81 Fetish 1 1.9
309.81 Post-traumatic Stress Disorder 1 1.9
Current Status of All (N = 53)
Total Revoked (n = 40)
Violent crime re-offense/revoke 15 28.3
Attempted murder 3 5.7
Assault with weapon 7 13.2
Arson 2 3.8
Sexual assault 2 3.8
Terrorist threat with weapon 1 1.9
Nonviolent crime/revoke 25 47.2
Nonviolent criminal re-offense 5 9.4
Drug/alcohol abuse violation 7 13.2
Treatment noncompliance/AWOL 10 18.9
Psychiatric decompensation 3 5.7
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 3 (continued)
Characteristic and category n % Range Mean SD
Total Success 13 24.5
Still in CONREP 10 18.9
Released from all supervision
in the Community 3 5.7
Time in CONREP until Current Date or
Revoke Date, by status group
Violent 15 2.0-84.0 40.9 27.2
Nonviolent 25 2.0-76.0 23.0 17.9
Success 13 12.0-79.0 45.9 18.9
Note. CONREP = Conditional Release Program of the California Department of Mental Health.
00
■vl
The term “instant crime” refers to the crime for which the patient was
committed to the state hospital. The instant crime distribution was as
follows: murder = 20, assault = 20, sexual offense = 4, arson = 2, robbery =
5, kidnapping = 1, theft of personal identity = 1 (total = 53).
The distribution of reasons for revocation was as follows: violent
revoked = 15 and nonviolent revoked = 25. The violent group were patients
who had reoffended violently, as follows: attempted murder = 3, assault
with a deadly weapon = 7, arson = 2, sexual assault = 2, and terrorist threat
with a weapon = 1. The nonviolent revoked group of 25 included nonviolent
crimes = 5, drug/alcohol abuse = 7, treatment noncompliance/AWOL = 10,
and psychiatric decompensation = 3. A total of 40 were revoked to the
hospital.
There were 13 successes (participants who remained in the com
munity without reoffending) in this study. Ten had remained successfully on
outpatient treatment, and 3 others been released by the court from
mandated treatment and were living in the community unsupervised.
Preliminary Group Comparisons
The means and standard deviations of each predictor for the three
outcome groups—success, violent reoffense, and nonviolent reoffense—
are shown in Table 4. The mean values of the scores on the violent and
nonviolent reoffense groups appear to be very close and f-test comparisons
found the differences to be nonsignificant, as shown in Table 5. Due to this
finding, the violent and nonviolent groups were combined as one
88
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4
Means and Standard Deviations of the Predictor Variables for Three
Outcome Groups
Outcome groups
Predictor Success Nonviolent Violent
variable M SD M SD M SD
Psychopathy (H7)
Psychosis (C3)
Substance Abuse (H5)
PCL:SV Total
HC-15 Total
0.85 0.69 1.36
1.08 0.86 1.08
1.77 0.60 1.80
13.85 4.51 15.92
16.38 4.93 20.16
0.81 1.53 0.74
0.70 1.13 0.64
0.50 1.73 0.70
5.91 17.33 4.73
6.15 20.80 4.38
Note. PCL:SV = Psychopathy Checklist: Screening Version; HC-15 = the
Historical, Clinical, and Risk Management-20 instrument, excluding the five
Risk Management items.
89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5
Results oft Tests of the Mean Differences Between Violent and Nonviolent
Re-Offense Groups for the Predictor Variables Psychopathy Psychopathy
Checklist: Screening Version Total Score, Historical, Clinical-15 Total
Score, and the Three Other Predictors Psychopathy, Active Psychosis, and
Substance Abuse
Predictor
Violent
(n = 15)
M SD
Nonviolent
(n = 25)
M SD df t
PCL:SV total 17.33 4.73 15.92 5.91 38 .786
HC-15 total 20.80 4.38 20.16 6.15 38 .352
Psychopathy (H7) 1.53 0.74 1.36 0.81 38 -.675
Active psychosis (C3) 1.13 0.64 1.08 0.70 38 -.240
Substance abuse
history (H5) 1.73 0.70 1.80 0.50 38 -.350
Note. PCL:SV = Psychopathy Checklist: Screening Version; HC-15 = the
Historical, Clinical, and Risk Management-20, excluding the five Risk
Management items; H7 = item 7 on the Historical scale of the Historical,
Clinical-15; C3 = item 3 on the Clinical scale of the Historical, Clinical-15;
H5 = item 5 on the Historical scale of the Historical, Clinical-15.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
group, labeled “all revoke.” The hypothesis testing was carried out using
these two outcome groups, success and all revoke.
Hypothesis 1
Hypothesis 1 stated: Violent and nonviolent reoffenders will have
higher psychopathy than nonreoffenders. A comparison of the means
between the success group and the all revoke group is shown in Table 6.
Table 6
Results of t Tests of the Mean Differences Between Successes and Violent/
Nonviolent Re-Offense Groups for the Predictor Variables Psychopathy
Psychopathy Checklist: Screening Version Total Score, Historical, Clinical-
15 Total Score, and the Three Other Predictors Psychopathy, Active Psy
chosis, and Substance Abuse
Predictor
Successes
(n = 13)
M SD
Violent and
Nonviolent
(n = 40)
M SD df t
PCL-.SV total 13.85 4.51 16.45 5.48 51 -1.55
HC-15 total 16.38 4.93 20.40 5.50 51 -2.34*
Psychopathy (H7) 0.85 0.69 1.43 0.78 51 -2.39*
Active psychosis (C3) 1.08 0.86 1.10 0.67 51 -0.10
Substance abuse
history (H5) 1.77 0.60 1.78 0.58 51 -0.03
Note. PCL:SV = Psychopathy Checklist: Screening Version; HC-15 = the
Historical, Clinical, and Risk Management-20, excluding the five Risk
Management items; H7 = item 7 on the Historical scale of the Historical,
Clinical-15; C3 = item 3 on the Clinical scale of the Historical, Clinical-15;
H5 = item 5 on the Historical scale of the Historical, Clinical-15.
*p< .05.
91
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 6 shows that the mean of the scores for the all revoke group
on psychopathy were significantly higher than the mean of the scores of the
success group. The p value of this comparison was .02, using a two-tailed
test. This finding is in accord with the hypothesized outcome in this study,
and this allows for a one-tailed test, which increases the significance to a
p value of .01. This suggests that Hypothesis 1 is true.
Hypotheses 2 and 3
Hypothesis 2 stated: Violent and nonviolent reoffenders will show
signs of active psychosis more than nonreoffenders. Table 6 shows that
the comparison between these two groups, success and the violent and
nonviolent reoffenders, on this variable was nonsignificant. This suggests
that Hypothesis 2 is false.
Hypothesis 3 stated: Violent and nonviolent reoffenders will have
more substance abuse problems than non reoffenders. Table 6 shows that
the comparison between these two groups, success and violent and non
violent reoffenders, on this variable was nonsignificant. This suggests that
Hypothesis 3 is false.
Hypothesis 4
Hypothesis 4 stated: The 15 items of the Historical and Clinical
scales of the HCR-20 will more accurately predict violent and nonviolent
reoffense in this population than the PCL.SV. Table 6 shows that the
violent and nonviolent reoffenders group scored significantly higher on the
HC-15 than the success group. This finding was a directionally expected
finding and allows for a one-tailed test, increasing the p value to .01. Table
92
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6 shows that the violent and nonviolent reoffenders group scored higher
than the success group on the PCL:SV and the difference in the means
approached significance at a p value of .06. In order to further compare the
predictive accuracy of the two assessment instruments (HC-15 and
PCL.SV) a receiver operating characteristic curve analysis was done.
To further examine the relationship between the two assessment
instruments an analysis was computed using the Receiver Operating
Characteristic (ROC) curve (Vida, 1993, 2001). The ROC shows the
relationship of true positives, false positives, true negatives, and false
negatives on a binomial graph as the “sensitivity” and “specificity” of the
two risk assessment tools.
Sensitivity is the number of true cases of a disease detected by
the test (true positives, TP) divided by the number of all diseased
subjects, whether detected by the test (true positive TP) or not
(false negatives, FN). Specificity is the number of true negatives
(nondiseased subjects who were considered negative by the test,
TN) divided by the number of all nondiseased subjects, whether
considered positive (false positive, FP) or negative (true negative,
TN) by the test. (Vida, 2001, p. 8)
A mathematical symbolization for the X and Y axes on the graph
would be:
TP
Sensitivity = TP + FN
TN
Specificity = FP + TN
The ROC curve is a graphic plot of “ sensitivity (%)” (True positive
fraction) on the Y-axis versus “1-specificity (%)” (False positive fraction) on
the X-axis produced by the range of scale scores or at various selected cut
off points. The larger the number of possible scores a scale has, the more
93
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
points will define the ROC curve. When the proportion of true positives is
the same as the proportion of false positives for each score, the ROC curve
is a diagonal line from the bottom left corner to the top right corner of the
graph. In this situation, the AUC will be .50, or 50%. The better a test is,
the closer is its ROC curve to the upper left hand corner of the graph, where
the true positive rate is 1 and the false positive rate is 0. A larger AUC
indicates better prediction. ROCs have been used in other research to
predict criminal and violent behavior (Gardner, Lidz, Mulvey, & Shaw, 1996;
Mossman, 1994; Serin & Lawson, 1987).
To compare these two instruments for predictive usefulness, the
AUC was calculated and compared (see Figure 1 and Table 7). The
PCL.SV score analysis shows that the all revoked offenders had an AUC of
.678, indicating a 67.8% chance that a patient who reoffends will obtain a
higher score on the PCL:SV than will a randomly chosen patient who does
not reoffend. The AUC for the HC-15 for the all revoked groups was .727,
indicating a 72.7% chance that a patient who reoffends will obtain a higher
score on the PCL:SV than will a randomly chosen patient who does not
reoffend. There was 5.9% difference in favor of the HC-15. This finding
suggests that Hypothesis 4 may be true and should not be rejected.
In summary, the results presented show that violent and nonviolent
reoffenders did significantly score higher on psychopathy than non
reoffenders in this study. However, signs of active psychosis and sub
stance abuse problems did not differentiate between the reoffenders and
the nonreoffenders in this study. The HC-15 assessment items more
accurately identified reoffenders in this study.
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ROC Curve
1.00
— -J
.75-
.50-
Source of the Curve
Total Historical,
Clinical 1 5 Score
>, .25-
■ A p *
> . .
" 45 ^
'5 5
c
£0 0.00. _
0.00
Total PCL:SV Score
.25 .75 1.00
1 - Specificity
Figure 1. The test result variable(s): Psychopathy Checklist: Screening
Version (PCL:SV) Total Score with Area Under the Curve (AUC) = .678;
and Historical, Clinical-15 (HC-15) Total Score with AUC = .727. The
crossing points of the graph lines indicate at least one tie between the
positive actual state group and the negative actual state group by the two
instruments. The AUC shows the percentage of accurate predictions of
violent and nonviolent reoffense by the PCLSV and the HC-15.
95
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 7
Coordinates of the Curve
Test result variable(s)
Positive if
greater than
or equal toa Sensitivity 1 - specificity
Pyschopathy Checklist 2.00 1.000 1.000
Total Sum 4.50 .950 1.000
6.50 .925 .923
8.00 .875 .846
9.50 .850 .846
11.00 .825 .769
12.50 .825 .615
13.50 .800 .538
14.50 .700 .462
15.50 .650 .462
16.50 .600 .231
17.50 .550 .154
18.50 .500 .154
19.50 .400 .154
20.50 .200 .077
21.50 .150 .000
22.50 .050 .000
24.00 .000 .000
Total Clinical and 6.00 1.000 1.000
Historical Score 7.50 .975 1.000
8.50 .950 .923
9.50 .950 .846
10.50 .925 .846
12.00 .900 .846
13.50 .900 .769
14.50 .825 .615
15.50 .825 .462
16.50 .775 .462
18.00 .725 .462
19.50 .575 .462
20.50 .525 .231
21.50 .500 .231
22.50 .475 .154
23.50 .350 .000
24.50 .325 .000
25.50 .150 .000
26.50 .075 .000
28.00 .025 .000
30.00 .000 .000
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 7 (continued)
Note. The test result variable(s): Psychopathy Checklist Total Sum, Total
Clinical and Historical Score has at least one tie between the positive actual
state group and the negative actual state group.
a The smallest cutoff value is the minimum observed test value minus 1, and
the largest cutoff value is the maximum observed test value plus 1. All of
the other cutoff value are the averages of two consecutive ordered
observed test values.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 5
DISCUSSION
This chapter contains a summary of the study, a discussion of the
findings related to each of the hypotheses, implications for explanatory
theories, implications for clinical practice, and recommendations for future
research.
Summary of the Study
The primary purpose of the study was to determine the predictive
value of the presence of psychopathy, active signs of psychosis, and
substance abuse problems with general and violent reoffense among
schizophrenic patients released from the state hospital as NGRI.
The secondary purpose was to determine whether dynamic or
changeable clinical factors, as assessed on the Clinical or “C” scale of the
HCR-20, improved the risk assessment of static or unchanging criminal
behavior history and durable psychopathic personality factors as assessed
by the PCL:SV.
In this study 40 of the 53 (75.5%) sample cases reoffended violently
or nonviolently and 13 (24.5%) reoffended violently. The total reoffense
rate was somewhat higher than the recidivism rates in the United States
reported by the BJS for the years 1983 and 1994. These rates were 62.5%
for 1983 and 67.5% for 1994 (Beck & Shipley, 1989; Langan & Levin,
2002).
98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Discussion of the Hypotheses
Four hypotheses were tested:
1. Violent and nonviolent reoffenders will have higher psychopathy
then nonreoffenders.
2. Violent and nonviolent reoffenders will show signs of active
psychosis more than nonreoffenders.
3. Violent and nonviolent reoffenders will have more substance
abuse problems than nonreoffenders.
4. The 15 items on the Historical and Clinical scales of the HCR-20
will more accurately predict violent and nonviolent reoffense in this
population than the PCL:SV.
In brief summary, the hypothesis that higher psychopathy would
predict future violent and nonviolent reoffense was supported at the p value
of .01 level of significance. The hypotheses that active signs of psychosis
and having more substance abuse problems would predict reoffense were
not supported and were nonsignificant. Both assessment instruments
predicted reoffense. The PCL:SV captured or correctly predicted 67.8% of
those who reoffended violently and nonviolently. The HC-15 scales of the
HCR-20 captured or correctly predicted 72.7% of those who reoffended
violently and nonviolently. These numbers show that the HC-15 items
better predicted reoffense than did the PCL:SV. The conclusions related to
each hypothesis are discussed below.
Hypothesis 1: Psychopathy
Hypothesis 1 stated that violent and nonviolent reoffenders would
have higher psychopathy than nonreoffenders. Analysis of the data from
99
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
this study sample supports this hypothesis at the p value of .01. Several
factors are worth noting.
First, this hypothesis was stated in a manner to express the
expected outcome. This expected outcome is in agreement with the pro
fessional observations of the author working with diagnosed schizophrenic
forensic patients for over 10 years and is in agreement with other
researchers (Cooke & Michie, 2001; Salekin et al., 1996). However, at
least one study reported that the presence of schizophrenia was a negative
predictor of reoffense (Harris et al., 1993; Lidz, Mulvey, & Gardner, 1993).
In the present study 40 of the 53 cases reviewed did reoffend and were
readmitted to the state hospital. It appears that assessing for the presence
of psychopathy is an important factor when considering individuals for
outpatient readiness.
Second, the way in which the study participants were chosen is
meaningful. Everyone in the CONREP population had criminal histories, by
legal definition of membership and by court commitment. Past criminal
behavior has been a known and established risk factor for future criminal
offense (Harris & Rice, 1997; Langan & Levin, 2002). So, it is not surprising
that a group of seriously mentally ill CONREP clients who had criminal
histories would also show higher scores on the psychopathy and would
reoffend at a high rate (75.5%). The actual psychopathy scores on the
PCL:SV among the 15 violent reoffenders was a mean of 17.3, but this
included one score of 6. The mean for the other 14 violent reoffenders was
18.1, which is slightly above the PCL:SV manual’s designated score of 18
for the presence of psychopathy. It appears that this study sample’s Violent
100
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Revoked group’s mean score was right on the dividing score between
“possible psychopathy” and “ serious psychopathy.” The scores of 13-17
(the “middle range”) have been historically less predictive of violent
outcome and reoffense in most other studies using the PCL-R or PCL:SV
(Hare, 1991; Hart et al., 1995). The possible lower overall mean of the
psychopathy scores in this sample may account for the lack of a differentia
tion between violent and nonviolent reoffenders. The cases reviewed in this
study may be similar to state hospital patients upon release, but they are
most probably not representative of non-state hospital patients.
The third factor that is that, in spite of the type of intensive interven
tions and supervision given in the CONREP program statewide, a large
number of patients reoffended, but did so nonviolently. This closely
monitored supervision most likely prevented some violence from occurring
among the other 25 patients who were revoked for nonviolent reasons.
CONREP supervision includes a minimum of three contacts per week
consisting of one individual therapy session, one group therapy session,
and one random drug screen (Church, 1991). Ninety percent of CONREP
patients begin with more than the minimum of staff-client contacts and
intervention opportunities. Frequent observations and frequent contact
afford opportunities to adjust medications quickly to avoid or prevent
psychiatric decompensation and/or aggressive behavior (Wiederanders &
Choate, 1994). This effectiveness seemed evident in that only 3 of the 40
revoked patients were revoked for psychiatric decompensation.
Fourth, the system of qualifying and obtaining outpatient status tends
to eliminate the highly sociopathic or psychopathic schizophrenic from
101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
hospital release. Patients who show strong psychopathic behaviors are the
ones who violate the unit rules, are assaultive and aggressive with others,
and are manipulative, while showing very little remorse for their behavior
and their victims. These qualities are negative indicators for the evaluators
when assessing patients for outpatient readiness. This aggressive behavior
must be controlled and eliminated in order for a patient to receive outpatient
treatment. This tendency to eliminate or preclude the highly psychopathic
NGRI patient from hospital release (Sewani Memo, CONREP staff,
personal communication, 2001) also reduces the number of high scorers in
this study.
In summary, psychopathy is an important risk factor of violent and
nonviolent reoffense in this population and the intensive case management
of and assessment for those patients with high scores on psychopathy
appears warranted. This said, the majority of reoffenses were nonviolent in
nature and may speak to the effectiveness of the CONREP program and
argue against the fear often aroused in the public’s mind about the schizo-
phrenically ill patients released from state hospitals.
Hypothesis 2: Active Psychosis
Hypothesis 2 stated that violent and nonviolent reoffenders would
show signs of active psychosis more than nonreoffenders. The analysis of
the data collected in this study did not support this hypothesis.
While it may appear as a given that a previously violent and actively
psychotic person is likely to become violent again (Monahan et al., 2001),
these two predictors were not correlated significantly with Violent Revoke or
102
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Nonviolent Revoke (or offense). In the past dozen years of research, the
presence of major mental illness has been shown to be a significant risk
factor for violent and aggressive behavior (Monahan et al.). Many studies
of psychiatric hospital admissions have recorded recent violence to have
been committed by those who are actively psychotic (Arango, Calcedo
Barba, Gonzalez-Salvador, & Calcedo Ordonez, 1999). Large epidemio
logical studies (e.g., Brennan, Mednick, & Hodgins, 2000; Swanson, Holzer,
Ganju, & Jono, 1990) have linked the presence of mental illness with
violence and aggression.
The CONREP program also frequently readmits/revokes a patient
from outpatient status and readmits the person into a state psychiatric
hospital when he or she becomes actively psychotic (Wiederanders &
Choate, 1994). This is done with the belief that psychotic behavior is a high
risk for violence and criminal offense. So, why was there so little to no
evidence that active psychosis predicts or is associated with future violence
and offense?
One plausible explanation centers on the fact that CONREP is very
effective in carrying out its mandate to monitor and supervise the patients
put in CONREP care. The CONREP patients are, in fact, treated pro
actively to prevent decompensation, aggression and violence, and rehos
pitalization as much as possible. The average CONREP patient in one
southern California CONREP clinic has about 4-10 hours per week of pro
fessional staff monitoring at the clinic and an added two to five face-to-face
contacts in home visits, collateral contacts, and medication reassessments
by the staff psychiatrist (Church, 1991). This intensive level of treatment
103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
has served the CONREP treatment milieu well in preventing psychiatric
decompensation. In this study, only 3 of the 25 nonviolent reoffenders were
revoked for psychiatric decompensation.
Yet another answer to the question of why active psychosis did not
show significance in this study may be, as some existing literature sug
gests, that active psychosis is not a predictor of violence and aggression
(Rasmussen & Levander, 1995). Those studies that have shown a link
between delusions and violence showed it for specific types of delusions,
termed TCO (Link & Stueve, 1994). This means that, when a person
suffering from active psychotic processes feels that he or she is in imminent
danger and acts violently or aggressively, he or she does so in order to
protect himself or herself from others. In this author’s 12 years of CONREP
experience, almost every CONREP case has been of this nature. In other
words, the initial crimes leading to mandated psychiatric treatment were
most often committed when the patient felt that he or she was in grave
danger while in delusional and paranoid mental states.
In most CONREP patients, this type of delusion is managed suc
cessfully with medication prior to release from the hospital and monitored
very carefully so that it does not reappear. It follows that, especially in the
months just prior to release from the hospital, symptoms of psychosis would
not be present. However, other types of “benign” hallucinations and
delusions tend to occur and would be coded as active psychosis in this
study sample. Perhaps, this explains the lack of a significant or consistent
finding regarding research that shows “ active psychosis” as a hypothesized
predictor of criminal offending.
104
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Hypothesis 3: Substance Abuse
Hypothesis 3 stated that violent and nonviolent reoffenders would
have more substance abuse problems than nonreoffenders. Substance
abuse history as a predictor was found to be nonsignificant in this set of
data. Several factors may explain this finding.
Hypothesis 3 stated that violent and nonviolent reoffenders would
have more substance abuse problems than nonreoffenders. Substance
abuse history as a predictor was found to be nonsignificant in this set of
data. The same rationale described above partially applies as a reason for
nonsignificance with this variable.
First, the relationship link between substance abuse and violence
may be a spurious one (Parker & Auerhahn, 1998; Rice & Harris, 1995),
meaning that there is no link between substance abuse and violence. This
issue continues to be debated in research literature. The meta-analysis of
60 controlled alcohol and violence studies by Bushman (1997) indicated no
evidence for the physiological disinhibition hypothesis to link alcohol and
violence. The same study also showed no evidence for another popular
model, alcohol-related expectancy. The causal model supported in this
meta-analysis was that alcohol “indirectly causes aggression by producing
changes within the person that increase the probability of aggression”
(p. 241). The term “indirectly” may be the key to be Bushman’s findings
versus other studies that found no relationship or link between alcohol and
aggression (Parker & Auerhahn).
Second, the lack of a link between substance abuse and violence
and reoffense may indicate an inadequate method of measuring the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
presence of substance abuse. In this study substance abuse was coded as
any past history of significant problematic substance abuse. The recency
or remoteness of the abuse was not coded and that information was not
available for coding.
In spite of this, the revocation of 7 of the reviewed persons who were
in the Nonviolent Revoked group was due to drug abuse. All of these had
drug abuse histories. But 7 other individuals among the Not Revoked group
also had drug abuse histories. These findings need further exploration and
further study. It seems obvious to conclude that, if a patient fails outpatient
treatment for abusing drugs, he or she will be likely to have abused drugs
before. However, past drug abuse in this study did not predict treatment
failure due to relapse to drug abuse.
Third, the prerelease selection criteria used to place people in the
outpatient program tend to exclude anyone with a recent substance abuse
relapse episode. This results in outpatient treatment of those patients
whose substance abuse behaviors have been in remission for longer
periods of time and those who are less likely to return quickly to their old
drug use behaviors.
Fourth, in post-release, CONREP monitors patients very carefully
and frequently performs weekly random drug screen tests to enforce this
stringent abstinence requirement. A violation of the abstinence term and
condition of treatment could result in the loss of freedom for the patient and
a return or revocation to inpatient treatment, thereby providing a strong
deterrent to relapse behavior.
106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The fifth factor that mediates against drug or alcohol abuse as a
predictor of violent and nonviolent offending is that all of the CONREP
clients have food and shelter guaranteed. There is no necessity to return to
the old drug lifestyle as a means of survival when all of the necessities are
already supplied (Akers, 1998).
Sixth, some who abused drugs in the past may have done so in an
attempt to medicate their symptoms (Khantzian, 1997). When a person is
on CONREP, their medications are supplied and maintained by court man
date, and self-medicating psychiatric symptoms are not a major motivation.
This is coupled with frequent medication visits and readjustments for any
negative psychotropic medication side effects as needed. This results in
less urgency or striving for the effect of drug use among the mentally ill.
Seventh, another mediating variable is CONREP’S common utiliza
tion of community treatment programs or augmented CONREP services to
help a person to overcome a substance abuse relapse episode and return
to abstinence. This means that many who have relapsed to drug abuse are
not revoked; instead, they are treated in the outpatient setting to overcome
the drug abuse without returning to the hospital. This practice shows that
the staff of the outpatient programs deems such relapse incidents to be
nondangerous or nonviolent risk behavior for these individuals.
Some studies show that drug abuse is associated with property
crimes among those who use in order that they may continue to purchase
their drug of choice (Goldstein, 1989). The violence associated with drug
abuse could be because of the financial commerce involved in drug dealing
more than from the biochemical effect of the drug itself. Some illicit drugs
107
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
have been found to be associated with increased violent behavior. These
include alcohol, cocaine, and PCP (Denison et al., 1997). Alcohol is often
associated with aggression in social situations, including with friends or
family (Wallace, 1989). There were too few cases of those revoked for
substance abuse (7) in this study to make a guess about the dynamics of
those identified as revoked for substance abuse.
Eighth, another factor that is described in many studies (Goldstein,
1989; Turner & Tsuang, 1990) about the relationship between substance
abuse and violence or crime is this that different drugs have different effects
on people, and the same drug has different effects on different people.
In other words, a drug’s effect varies widely on the individual user and
between users. It is suggested that studies be conducted on the behavior
of persons who abuse each different drug. Often, an unaccounted-for
factor is that many people abuse many different drugs, either simul
taneously or serially. These people are often called polydrug users.
Polydrug abuse complicates the picture even more. In this study there
were both polydrug abusers and those who abused only their drug of
choice. The number of polydrug abusers by history was the largest group
(n = 18).
Further studies are needed to separate substance abusers by the
drug of choice. This would be helpful in trying to understand this complex
problem, which involves biological and social/behavioral effects for each
illicit substance.
108
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Hypothesis 4: Comparison of Predictive
Power of the Instruments
Hypothesis 4 stated that the 15 items on the Historical and Clinical
scales of the HCR-20 would more accurately predict violent and nonviolent
reoffense in this population than the PCL:SV. The t test reported in Table 6
shows that the violent and nonviolent reoffenders did have significantly
higher scores on the HC-15 item assessment at a p value of .01.
Table 6 f-test results using scores on the PCL:SV show the analysis
results for this instrument at a p value of .06. These results would indicate
that the HC-15 assessment was more predictive of the outcome in this
study.
Further comparison analysis of these two assessment instruments
using the ROC analysis shows a similar difference between the two (see
Figure 1). The PCL:SV identifies 67.8% of the violent and nonviolent re
offenders in the AUC, while the HC-15 identifies 72.7% of the violent and
nonviolent reoffenders in the AUC.
The data in this study showed the HC-15 to be slightly better in risk
assessment. Why? Immediately, a puzzling question arises. It is the
question about the large and robust literature on the PCL:SV and the
PCL-R that shows that, in prison and psychiatric populations, psychopathy,
as measured on the PCL-R/PCLSV, predict violent offenses quite well
(Hart et al., 1995).
The answer may be from the evidence that the lower or midrange
scores on the PCL instruments do not predict well the future or past violent
behavior of the individual in question (Hare, 1991; Hare et al., 1995). In
this study, the mean score of the violent revokees was 17.5. This included
109
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
one score of 6; it this score is removed from the sample, the mean score
for the 14 other violent reoffenders is raised to 18.1. While this is in the
range of the scores for signifying the presence of psychopathy, it is barely
so.
It may be that, in this sample of subjects, there were not enough high
scorers on the PCL:SV to show a strong correlation or prediction of violent
offending. It may also be the case that schizophrenic patients have a
different cut-off score to indicate psychopathy, or it may mean that in this
sample the PCLSV did not discriminate well the future behavior of psychi
atric patients with schizophrenia who were recommended for CONREP
outpatient treatment.
One possibility for the HCR-20/HC-15’s slightly better overall per
formance may be that the C scale of the HCR-20 focuses more on the
current or recent mental illness issues that impinge on current and future
behavior of forensically committed mentally ill patients. The epidemiological
studies cited earlier (e.g., Monahan et al., 2001) support a raised risk of
violence among the mentally ill, especially when substance abuse is
present. A difference was that the substance abuse was assessed as more
current in those in the epidemiological studies than in those who were
assessed in this study at the time just prior to obtaining outpatient treatment
status. In the post hoc analysis (see Tables 8 and 9), the Clinical items
scale showed more significance to outcomes than did the Historical items
scale.
110
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 8
Means and Standard Deviations for Violent and Nonviolent Groups (N = 25)
Item n Min Max Mean SD
PCL1 = Superficial 40 0 2 1.05 0.78
PCL2 = Grandiose 40 0 2 1.23 0.86
PCL3 = Deceitful 40 0 2 1.03 0.80
PCL4 = Lacks Remorse 40 0 2 1.50 0.64
PCL5 = Lacks Empathy 40 0 2 1.55 0.64
PCL6 = Doesn’t Accept Responsibility 40 0 2 1.62 0.67
PCL7 = Impulsive 40 0 2 1.62 0.63
PCL8 = Poor Behavioral Controls 40 0 2 1.62 0.54
PCL9 = Lacks Goals 40 0 2 1.55 0.60
PCL10 = Irresponsible 40 0 2 1.38 0.77
PCL11 = Adolescent Antisocial Behavior 40 0 2 1.02 0.86
PCL12 = Adult Antisocial Behavior 40 0 2 1.25 0.74
PCLSUM1 = sum of items 1-6 40 0 12 8.00 3.19
PCLSUM2 = sum of items 7-12 40 1 12 8.45 2.81
PCL total = sum of items 1-12 40 3 23 16.45 5.48
H1 = Previous Violence 40 0 2 1.77 0.58
H2 = Young Age at First Violent Incident 40 0 2 1.05 0.71
H3 = Relationship Instability 40 0 2 1.60 0.59
H4 = Employment Problems 40 0 2 1.40 0.74
H5 = Substance Use Problems 40 0 2 1.78 0.58
H6 = Major Mental Illness 40 2 2 2.00 0.00
H7 = Psychopathy 40 0 2 1.42 0.78
H8 = Early Maladjustment 40 0 2 1.53 0.64
H9 = Personality Disorder 40 0 2 1.08 0.94
H10 = Prior Supervision Failure 40 0 2 1.25 0.90
Total Historical = sum of items H1-H10 40 4 20 14.88 4.11
C1 = Lack Insight 40 0 2 1.50 0.68
C2 = Negative Attitudes 40 0 2 1.03 0.83
C3 = Active Symptoms of Major 40 0 2 1.10 0.67
Mental Illness
C4 = Impulsivity 40 0 2 1.13 0.72
C5 = Unresponsive to Treatment 40 0 2 0.75 0.78
Total Clinical = sum of items C1-C10 40 0 10 5.16 2.97
Total Clinical + Total Historical 40 7 29 20.40 5.50
PCL + Historical + Clinical (all items) 40 10 49 36.85 10.38
Valid N (listwise) 40
Note. PCL = Psychopathy Checklist; H = Historical and C = Clinical scales
of the Historical, Clinical-15.
111
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 9
Means and Standard Deviations for Success Group (N = 13)
Item n Min Max Mean SD
PCL1 = Superficial 13 0 2 1.08 0.76
PCL2 = Grandiose 13 0 2 0.69 0.85
PCL3 = Deceitful 13 0 2 0.69 0.75
PCL4 = Lacks Remorse 13 0 2 1.15 0.90
PCL5 = Lacks Empathy 13 1 2 1.46 0.52
PCL6 = Doesn’t Accept Responsibility 13 0 2 1.31 0.63
PCL7 = Impulsive 13 0 2 1.46 0.66
PCL8 = Poor Behavioral Controls 13 0 2 1.38 0.77
PCL9 = Lacks Goals 13 0 2 1.38 0.77
PCL10 = Irresponsible 13 0 2 1.38 0.65
PCL11 = Adolescent Antisocial Behavior 13 0 2 0.92 0.64
PCL12 = Adult Antisocial Behavior 13 0 2 0.92 0.76
PCLSUM1 = sum of items 1-6 13 3 11 6.38 2.63
PCLSUM2 = sum of items 7-12 13 3 11 7.46 2.44
PCL total = sum of items 1-12 13 6 21 13.85 4.51
H1 = Previous Violence 13 0 2 1.69 0.75
H2 = Young Age at First Violent Incident 13 0 2 0.69 0.85
H3 = Relationship Instability 13 0 2 1.08 0.76
H4 = Employment Problems 13 0 2 1.23 0.83
H5 = Substance Use Problems 13 0 2 1.77 0.60
H6 = Major Mental Illness 13 2 2 2.00 0.00
H7 = Psychopathy 13 0 2 0.85 0.69
H8 = Early Maladjustment 13 0 2 1.46 0.78
H9 = Personality Disorder 13 0 2 0.85 0.90
H10 = Prior Supervision Failure 13 0 2 0.85 0.80
Total Historical = sum of items H1-H10 13 6 17 12.46 3.36
C1 = Lack Insight 13 1 2 1.38 0.51
C2 = Negative Attitudes 13 0 2 0.31 0.63
C3 = Active Symptoms of Major 13 0 2 1.08 0.86
Mental Illness
C4 = Impulsivity 13 0 1 0.46 0.52
C5 = Unresponsive to Treatment 13 0 2 0.69 0.75
Total Clinical = sum of items C1-C10 13 1 8 3.92 2.40
Total Clinical + Total Historical 13 8 23 16.38 4.93
PCL + Historical + Clinical (all items) 13 15 43 30.23 8.52
Valid N (listwise) 13
Note. PCL = Psychopathy Checklist; H = Historical and C = Clinical scales
of the Historical, Clinical-15.
112
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Implications for Explanatory Theories
Criminologists, psychologists, and medical scientists have generated
many theories of the pathways to violent and criminal behavior. These
theories and theorists sometimes describe similar phenomena using differ
ent language or vocabularies.
More attention should be given to exploring the popular biopsycho
social model of medical and psychiatric treatment settings. The parameters
in each of the three fields— biology, psychology, and sociology— require
definition and clarity regarding how each field overlaps or supports or does
not support the other two fields.
Akers (1998) has written recently about social learning theory and
deviant behavior and is an example of the blending of the history of
criminology theory and psychology’s social learning theory. It would be
helpful to incorporate neurobiology and neuropsychiatry into Akers’s social
learning model. This is especially appropriate, given the growing research
on bioneurology and brain chemistry. The successes of antipsychotic
medications would be enhanced with comprehensive psychological and
sociological interventions for mentally ill individuals at risk for further
criminal activity (Harris & Rice, 1997; Kopelowicz, Liberman, & Zarate,
2002; Sutkers & Adams, 2001). This cooperative blending of theories using
the biopsychosocial model may move ahead the violence prevention hope
of society, especially among the chronically mentally ill.
The implication for the theory or hypothesis that the presence of
psychopathy predicts violent and nonviolent behavior was mildly supported
113
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and corroborates prior research with criminal offenders and with psychiatric
offenders (Rice & Harris, 1992,1995) and specifically schizophrenic
offenders (McNiel, 1994). That the correlation was higher for violent
offense than for nonviolent offense is also similar to prior research findings
(Rice & Harris; Valliant, Gristey, Pottier & Kosmyna, 1999). The “mildness”
of the correlation may also give support to the finding by Harris et al. (1993)
that schizophrenia served as a negative predictor for reoffense. The lack of
a more robust correlation between psychopathy and violence may be the
result of some of the limitations present in this study, such as lack of a con
trol group, bias in the participant selection process, and the small sample
size. Nevertheless, even with these limitations, the finding of a relationship
at the .01 level shows how strong the psychopathy construct operates in the
prediction and identification of those at high risk for both violent and non
violent reoffense in the schizophrenic population of this study.
The implication for the biopsychosocial model of schizophrenic
offenders is that personality disorders and psychopathy should occupy an
important part in the risk assessment process for the mentally ill who are
reviewed for release into the community. Assessment for psychopathy is at
least as important as the assessment of clinical behaviors, mental status,
and substance abuse behaviors.
The theory or hypothesis that active psychotic symptoms predict
violence was not supported in this study. The implication of this finding
should include that further study is needed at the minimum. In those further
studies the collection of information should include recent as well as past
and present specific nature of the symptoms of active psychosis, especially
114
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
since schizophrenic symptoms are episodic in nature (Link & Stueve, 1994;
McNiel, 1994). Documentation in research studies regarding the interven
tions made over a period of time to control active symptoms of psychosis
may also give more clarity to how and when active psychotic symptoms are
or are not significant risk factors for violent and nonviolent reoffense
(Mulvey & Lidz, 1993; Wiederanders & Choate, 1994).
The theory or hypothesis that substance abuse, especially substance
abuse history, predicts violent and nonviolent reoffense was not supported
in this study. However, the presence of current substance abuse was not
specifically and directly assessed. The seven nonviolent offenses for sub
stance abuse were represented by individuals who had a substance abuse
history. These 7 reoffenders comprised 31 % of the total of 19 substance
abusers in the total study (24 minus 5 caffeine abusers = 19). The 31% in
this study is about half of the offense rate (64%) found among those who
were dually diagnosed in the MacArthur Risk Assessment Study (Monahan
& Steadman, 1994; Monahan et al., 2001). This significant reduction in
reoffense rate found in this study could well be attributed to the close
monitoring that conditionally released NGRI offenders receive in the
CONREP program in California (Wiederanders & Choate, 1994).
The nonsignificant correlation found between substance abuse and
reoffense in this study lends more support to the “ spurious” hypothesis,
which states that, when violence and substance abuse co-occur, there is
another or spurious factor that accounts for the violence or deviance. The
discussions above point to psychopathy as the likely other or spurious
115
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
factor. Indeed, other researchers have found a co-occurrence between
psychopathy and substance abuse (Swanson, 1994) in deviant populations.
The theory or hypothesis that current dynamic, clinical variables,
such as those included in the HCR-20 instrument, add more accuracy to
risk assessments of purely historical or static factors was supported in this
study. Identifying and coding the presence of signs of active psychosis was
not significant in differentiating between offenders and nonreoffenders.
Merely identifying the presence of hallucinations and delusions of schizo
phrenic patients may not capture enough of the quality and quantity of an
active psychosis condition that predicts or leads to violent or nonviolent
offenses. Link and Stueve (1994) used 13 descriptions of delusional and
hallucinatory thoughts, the TCO model, to discover the amount of and just
what parts of active psychotic experience are linked with aggression and
violence. One example of these 13 items was “How often have you felt that
there were people who wished to do you harm?” Another example was
“How often have you felt that your mind was dominated by forces beyond
your control?” As more symptoms were disclosed, the amount and rate of
violent behavior rose. The TCO model shows some utility in differentiating
types and amounts of active thought disorders that lead to aggression and
violent behavior.
This study did not identify what types of delusions or hallucinations
were present in the cases reviewed. Often, it is just such positive symp
toms of schizophrenia that receive the most focus during many outpatient
readiness evaluations (Wiederanders & Choate, 1994). More emphasis
may need to be given to assessing the quality of the delusions and
116
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
hallucinations experienced by the patient being assessed, as was done in
the research by Link and Stueve (1994).
While there was no clear “ winner” between the PCL:SV and the
HCR-20, the PCL:SV showed the expected increasing progression in the
mean scores of those in the success group (13.85), to nonviolent offense
(15.92), to violent offense (17.33). The HC-15 mean scores also increased
in progression from success (16.38) to nonviolent offense (20.16) to violent
offense (20.80; see Table 4). The PCL:SV better identified the violent
versus the success groups. The HC-15 did not differentiate nonviolent and
violent offense groups but it differentiated the success group from all
reoffenders. This difference serves to show how using more than one
means of risk assessment (multimodal assessment) offers more information
that a single risk assessment instrument.
Implications for Clinical Practice
The surprising finding in this study was that the two items of the
Historical Clinical-15— itemH5 “Substance Abuse History” and item C3
“Signs of Active Psychosis” —were not significant in predicting reoffending
behaviors. What may be said about this sample population is that they did
again what they had done before, and it was not due to drug abuse or
symptoms of psychotic mental illness. This said, however, it remains that
monitoring and treating mental illness symptoms and substance abuse
problems are necessary with CONREP patients with schizophrenia. The
existing literature cited herein certainly supports monitoring for all three of
the hypothesized predictors of risk for reoffense: psychopathy, active
117
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
psychosis, and substance abuse problems (De Pauw & Szulecka, 1998;
Link & Stueve, 1994; Steadman et al., 1998; Swanson, 1994; Swanson,
Borum etal., 1996).
The above discussion leads to the logical steps of implementing
psychopathy assessment in all outpatient readiness evaluations for this
population. It appears useful to employ multiple risk assessment tools in
the evaluation process. The assessment process should include a
“ criminogenic needs” assessment and be followed up with treatment plans
that address those needs in close supervision (Hanson & Harris, 2000).
One such assessment tool is the LSI-R (Andrews & Bonta, 1995). When
addressing the needs of sexual offenders, sexual deviance should be
assessed (Harris & Rice, 1997; Serin, Mailloux, & Malcolm, 2001; Seto &
Barbaree, 1999). A phalometric assessment utilizing a variety stimulus
pictures has been used. Close supervision and some isolation from the
most helpless of prospective victims are essential in protecting the com
munity (Hanson & Harris). Family members and friends—those most
frequently victimized by the mentally ill—should be included in the biopsy
chosocial model of treatment rendered to the mentally ill offenders
(Kapelowicz et al., 2002).
Even more urgency is apparent for assessing each patient’s history
through criminal records, psychological reports, social histories, and
collateral informants. Face-to-face interviews should be preceded by a
careful record review so that important information is not overlooked. The
most dangerous patients may gloss over very important information from
their histories. Personality characteristics (Axis II), and particularly those
118
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
characteristics that mark the presence of psychopathy, must be assessed in
all psychiatric offenders.
Limitations of the Study
Regarding methodological limitations, the study is nonexperimental
in design. Caution should be used when making inferences related to
causation. The concerns of the study are of a public health nature. This
was an epidemiological study of the behavioral consequences of the
assessed personality and behavioral measures, psychopathy, active
psychosis, and substance abuse prior to hospital release. The conse
quence measured was violent and nonviolent reoffense.
The relatively small number of available participant records (N = 53)
limits the power of analysis to find or confirm differences between groups.
Any finding of significance or nearing significance should be considered
tentative, as so few patient records were available for review in each group:
Success (did not reoffend) = 13, Violent and Nonviolent (did reoffend) = 40.
This relatively small number of patient charts was not an intentional limit in
the design. At the outset, many more charts were reported to be available
in the original database and feasibility investigation. As the project
progressed, several governmental, institutional, and administrative limits
made a large number of records unavailable for review. During the actual
work of reviewing the medical records, several more unforeseen problems
arose, including deletion and destruction of records due to storage age,
errors in diagnoses, and errors in the recorded age of the patient. Another
problem that eliminated many records from review was that a group of
119
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
eligible patient records had been moved when the patients were transferred
to another hospital.
Regarding procedural criteria, limiting the study to patients with a
diagnosis of schizophrenia reduced the number of files available for review.
This limit was set in order to compare among schizophrenics only and not
to confuse other types of major mental illness with these specific measures.
Many other studies include mixed types of mental illness diagnoses. It was
thought that, since schizophrenia is the major diagnosis among this forensic
population, this diagnostic group warranted its own comparison study.
Age was a limiting constraint. It is known that young age is a strong
predictive factor among those who commit crimes. Age is also a factor in
the onset and the diagnosis of schizophrenia. Initially, the age limit was set
at 25-35 years; this age limit was deemed appropriate because most
schizophrenia is diagnosed during that age range and most criminal
behavior declines after age 40. In the end, the age range included those
who were 23 to 38 years old. Several patient charts were at first identified
for inclusion, only to be rejected later when it was discovered that the
patient was too young or too old to meet the age criterion.
This age limit factor eliminated an older age group of patient records.
It was found that the average age of those admitted to CONREP from state
hospitals was 38.5 years. This fact may limit the usefulness of applying the
results of this study to the CONREP population, since the average age in
this study is much younger, about 31 years. This study of younger
CONREP patients may be used in comparison with other studies of older
CONREP patients in the future.
120
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Another limiting factor is that all of the participants were chosen from
among state hospital forensic files. This means that all of the participants
had criminal histories; indeed, all but one were violent past offenses. Prior
criminal history is a major factor in assessing future criminal offense risk.
The participants were all at higher risk for reoffense than the general
population or even the population of schizophrenic patients that do not have
a criminal history.
Because of the above limitations, this study should not be used for
generalization to other outpatient populations of schizophrenic individuals.
Further study is needed to understand the risk factors of violent and
nonviolent offenses among schizophrenic persons living in the community.
Recommendations for Future Research
This study was launched to find predictors of violent and nonviolent
reoffense and revocation among the population of schizophrenics who were
found not guilty by reason of insanity and who had committed felonies,
many of them violent. Three variables were presented as possible pre
dictors and two instruments were utilized to code these three variables, as
well as others. The two instruments, HCR-20/HC-15 and PCL:SV, were
also part of the study to see which (or whether either) would be useful in
making reoffense predictions with this population. The two instruments
were found to be predictive of reoffense, and the HC-15 (with the five R
items excluded) was found to be slightly more accurate as a predictor with
the ROC analysis (see Figure 1 and Table 7). Psychopathy significantly
121
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
correlated with reoffense; active psychosis and substance abuse problems
did not significantly correlate with reoffense.
A replication of this study is recommended with the following
improvements: (a) Use a much larger sample, about 200-300 cases;
(b) select several sites or hospitals; (c) broaden the age range to include
older patients, since the average age of the NGRI patient released from
state hospitals is about 38 years; (d) include more than NGRI legal status,
especially including PC 2962 patients, Mentally Disordered Offenders; and
(e) keep a homogeneous diagnosis group, which may continue to be useful
for clinical applications.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
REFERENCES
123
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
REFERENCES
Abram, K. M. (1989). The effect of co-occurring disorders on criminal
careers: Interaction of antisocial personality, alcoholism, and drug
disorders. International Journal of Law and Psychiatry, 72,133-148.
Akers, R. L. (1985). Deviant behavior: A social learning approach (3rd
ed.). Belmont, CA: Wadsworth.
Akers, R. L. (1998). Social learning and social structure: A general theory
of crime and deviance. Boston: Northeastern University Press.
American Psychiatric Association. (1974). Clinical aspects of the violent
individual. Washington, DC: Author.
American Psychiatric Association. (1994). Diagnostic and statistical man
ual of mental disorders (4th ed.). Washington, DC: Author.
Andreason, N. C., & Black, D. W. (2001). Introductory textbook of psy
chiatry. Washington, DC: American Psychiatric Publishing.
Andrews, D. A., & Bonta, J. (1995). The Level of Service inventory, Re
vised. Toronto: Multi-Health Systems.
Arango, C., Calcedo Barba, A., Gonzalez-Salvador, R., Calcedo Ordonez,
A. (1999). Violence in inpatients with schizophrenia: A prospective
study. Schizophrenia Bulletin, 25, 493-503.
Bandura, A., & Walters, R. H. (1963). Social learning and personality de
velopment. London, England: Holt, Rinehart and Winston.
Bartels, S. J., Drake, R. E., Wallach, M. A., & Freeman, D. H. (1991).
Characteristic hostility in schizophrenic outpatients. Schizophrenia
Bulletin, 17, 163-171.
Beck, A. J., & Shipley, B. E. (1989). Bureau of Justice Statistics special
report: Recidivism of prisoners released in 1983. Washington, DC:
U.S. Department of Justice, Office of Justice Programs.
Blomhoff, S., Seim, S., & Friis, S. (1990). Can predictions of violence
among psychiatric inpatients be improved? Hospital and Community
Psychiatry, 41, 771-775.
Bonta, J., Harman, W. G., Hann, R. G., & Cormier, R. B. (1996). The pre
diction of recidivism among federally sentenced offenders: A re-
evaluation of the SIR scale. Canadian Journal of Criminology, 38,
61-79.
124
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Bonta, J., Law, M., & Hanson, K. (1998). The prediction of criminal and
violent recidivism among mentally disordered offenders: A meta
analysis. Psychological Bulletin, 123, 123-142.
Borom, R. (1996). Improving the clinical practice of violence risk assess
ment: Technology, guidelines, and training. American Psychologist,
51, 945-956.
Brennan, P. A., Mednick, S. A., & Hodgins, S. (2000). Major mental dis
orders and criminal violence in a Danish birth cohort. Archives of
General Psychiatry, 57, 494-500.
Brent, D. A., Johnson, B. A., Perper, J. Connolly, J., Bridge, J., Bartle, S., et
al. (1994). Personality disorder, personality traits, impulsive vio
lence, and completed suicide in adolescents. Journal of the Ameri
can Academy of Child and Adolescent Psychiatry, 3 3 ,1080-1086.
Burgess, E. W. (1968). Factors determining success or failure on parole.
In A. A. Bruce, A. J. Harno, E. W. Burgess, & J. Landesco (Eds.),
The workings of the indeterminate-sentence law and the parole sys
tem in Illinois (pp. 205-234). Montclair, NJ: Patterson Smith. (Origi
nal published 1928 by the State of Illinois)
Bushman, B. J. (1997). Effects of alcohol on human aggression: Validity
of proposed explanations. In M. Galanter (Ed.), Recent develop
ments in alcoholism. Vol. 13: Alcohol and violence: Epidemiology,
neurobiology, psychology, family issues (pp. 227-243). New York:
Plenum Press.
Castle, D. J., Phelan, M., Wessely, S., & Murray, R. M. (1994). Which pa
tients with non-affective functional psychosis are not admitted at first
psychiatric contact? British Journal of Psychiatry, 165,101 -106.
Church, M. F. (Ed.). (1991). Forensic conditional release program policiy
and procedure manual. Sacramento, CA: Department of Mental
Health, Office of Forensic Services.
Cleckley, H. (1982). The mask of sanity (rev. ed.). St. Louis, MO: Mosby
Cleckley, H. M. (1941). The mask of sanity (1sted.). St. Louis, MO:
Mosby.
Cloninger, C. R., Reich, T., & Guze, S. B. (1975). The multifactorial model
of disease transmission: Sex differences in the familial transmission
of sociopathy (antisocial personality). British Journal of Psychiatry,
127,11-22.
125
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Cloninger, C. R., Sigvardsson, S., Bohman, M., & von Knorring, A. L.
(1982). Predisposition to petty criminality in Swedish adoptees:
Cross-fostering analysis of gene-environmental interactions.
Archives of General Psychiatry, 3 9 ,1242-1247.
Cloninger, D. (1987). A systematic method for clinical description and
classification of personality variants: A proposal. Archives of Gen
eral Psychiatry, 44, 573-588.
Compton, W. W., Helzer, J. E., Hwu, H. G., Yeh, E. K., McEvoy, L., & Tipp,
J. E. (1991). New methods in cross-cultural psychiatry: Psychiatric
illness in Taiwan and the United States. American Journal of Psy
chiatry, 148, 1697-1704.
Cooke, D. J. (1995). Psychopathic disturbance in the Scottish prison
population: Cross-cultural generalizability of the Hare Psychopathy
Checklist. Psychology, Crime and Law, 2 , 101-118.
Cooke, D. J., Forth, A. E., & Hare, R. D. (Eds.). (1998). Psychopathy:
Theory, research, and implications for society. Dordrecht, The Neth
erlands: Kluwer.
Cooke, D. J., Kosson, D. S., & Michie, C. (2001). Psychopathy and ethnic
ity: Structural, item, and test generalizability of the Psychopathy
Checklist-Revised (PCL-R) in Caucasian and African American par
ticipants. Psychological Assessment, 13, 531-542.
Cooke, D. J., & Michie, C. (1997). An Item Response Theory analysis of
the Hare Psychopathy Checklist-Revised. Psychological Assess
ment, 9(3), 3-14.
Cooke, D. J., & Michie, C. (2001). Refining the construct of psychopathy:
Towards a hierarchical model. Psychological Assessment, 13(2),
171-188.
Cooke, D. J., Michie, C., Hart, S. D., & Hare, R. D. (1999). Evaluating the
Screening Version of the Hare Psychopathy Checklist, Revised
(PCLSV): An Item response theory analysis. Psychological As
sessment, 11, 3-13.
Cooper, H., & Hedges, L. V. (1994). The handbook of research synthesis.
New York: Russell Sage Foundation.
Dalton, K. S. (2000). Factors predicting recidivism among adult offenders
in the San Bernardino County correctional system. Dissertation,
Loma Linda University, Loma Linda, CA.
Darke, S., Kaye, S., Finlay-Jones, R., & Hall, W. (1998). Factor structure
of psychopathy among methadone maintenance patients. Journal of
Personality Disorders, 12,162-171.
126
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
De Pauw, K. W., & Szulecka, T. K. (1998). Dangerous delusions: Vio
lence and the misidentification syndromes. British Journal of Psy
chiatry, 152, 91-96.
Dean, C. W., Brame, R., & Piquero, A. R. (1996). Criminal propensities,
discrete groups of offenders, and persistence in crime. Criminology,
34, 547-574.
Denison, M. E., Paredes, A., & Booth, J. B. (1997). Alcohol and cocaine
interactions and aggressive behaviors. In M. Galanter (Ed.), Recent
developments in alcoholism. Vol. 13: Alcohol and violence: Epide
miology, neurology, psychology family issues (pp. 283-303). New
York: Plenum Press.
Dilts, S. L., Jr. (2001). Models of the mind: A framework for biopsycho
social psychiatry. Philadelphia: Brunner-Routledge.
Dohrenwend, B. P., Shrout, P., Link, B., Martin, J., & Skodal, A. (1986).
Overview and initial results from a risk factor study of depression and
schizophrenia. In J. E. Barret & R. M. Rose (Eds.), Mental disorders
in the community (pp. 184-215). New York: Guilford.
Doren, D. M. (1987). Understanding and treating the psychopath. New
York: Wiley.
Doren, D. M. (1996). Understanding and treating the psychopath (2nd
ed.). Northvale, NJ: Jason Aronson.
Douglas, K. S. (1996). Assessing the risk of violence in psychiatric outpa
tients: The predictive validity of the HCR-20 risk assessment
scheme. Unpublished master’s thesis, Simon Fraser University,
Burnaby, British Columbia, Canada.
Douglas, K. S., & Hart, S. D. (1996, March). Major mental disorder and
violent behavior: A meta-analysis of study characteristics and sub
stantive factors influencing effect size. Poster presented at the bien
nial meeting of the American Psychology-Law Society, Hilton Head,
SC.
Douglas, K. S., Ogloff, J. R. P., Grant, I., & Nicholls, T. L. (1999). Assess
ing risk for violence among psychiatric patients: The HCR-20 Vio
lence Risk Assessment Scheme and the Psychopathy Checklist:
Screening Version. Journal of Consulting and Clinical Psychology,
67, 917-930.
Douglas, K. S., Ogloff, J. R. P., & Nicholls, T. L. (1997). The assessment
of risk for violence in psychiatric inpatients [Abstract]. Canadian
Psychology, 38(2a), 111.
127
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Douglas, K. S., Webster, C. D., & Wintrup, A. (1996, August). The HCR-20
risk assessment scheme: Psychometric properties in two samples.
Paper presented at the annual conference of the American Psycho
logical Association, Toronto.
Drake, R. E., Osher, F. C., & Wallach, M. A. (1989). Alcohol use and
abuse in schizophrenia: A prospective community study. Journal of
Nervous and Mental Disease, 177, 408-414.
Engel, G. L. (1977). The need for a new medical model: A challenge for
biomedicine. Science, 196,129-136.
Engel, G. L. (1980). The clinical application of the biopsychosocial model.
American Journal of Psychiatry, 137, 535-544.
Eysenck, H. J. (1964). Crime and personality. London: Methuen.
Eysenck, H. J. (1977). Crime and personality (3rd ed.). London:
Routledge and Kegan Paul.
Eysenck, H. J., & Eysenck, S. B. G. (1978). Psychopathy, personality, and
genetics. In R. D. Hare & D. Schalling (Eds.), Psychopathic behav
ior: Approaches to research. Chichester, England: Wiley.
Fagan, J. (1993, Winter). Interaction among drugs, alcohol, and violence.
Health Affairs, 65-77.
Fagan, J., & Chin, K. (1990). Violence as regulation and social control in
the distribution of crack. In M. de la Rosa (Ed.), Drugs and violence:
Causes, correlates and consequences (pp. 8-43). Washington, DC:
NIDA.
Fagan, J., & Wilkinson, D. L. (1998). Social contexts and functions of
adolescent violence. In D. S. Elliott, B. A. Hamberg, & K. R. Williams
(Eds.), Violence in American schools: A new perspective (pp. 55-
93). New York: Cambridge University Press.
Farrington, D. P. (1977). Crime and personality (3rd ed.). St. Albans, NY:
Paladin.
Fogel, B. S., Schiffer, R. B., & Rao, S. M. (Eds.). (2000). Synopsis of neu
ropsychiatry. Philadelphia: Lippincott, Williams and Wilkins.
Forth, A. E., Hart, S. D., & Hare, R. D. (1990). Assessment of psychopathy
in male young offenders. Psychological Assessment: A Journal of
Consulting and Clinical Psychology, 2, 342-344.
Fulwiler, C., Grossman, H., Forbes, C., & Ruthazer, R. (1997). Early-onset
substance abuse and community violence by outpatients with
chronic mental illness. Psychiatric Services, 4 8 ,1181-1185.
128
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Gardner, W., Lidz, C. W., Mulvey, E. P., & Shaw, E. C. (1996). Clinical
versus actuarial predictions of violence of patients with mental ill
nesses. Journal of Consulting and Clinical Psychology, 64, 602-609.
Gendreau, P., Little, T., & Goggin, C. (1996). A meta-analysis of the pre
dictors of adult offender recidivism: What works! Criminology, 34,
575-607.
Gibbons, D. C. (1994). Talking about crime and criminals. Englewood
Cliffs, NJ: Prentice-Hall.
Goldstein, P. J. (1989). Drugs and violent crime. In N. A. Weiner & M. E.
Wolfgang (Eds.), Pathways to criminal violence (pp. 16-48). London:
Sage.
Goldstein, P. J. (1995). The drugs/violence nexus: A tripartite conceptual
framework. Drug Issues, 15, 493-506.
Gondolf, E. W., & White, R. J. (2001). Batterer program participants who
repeatedly reassault. Journal of Interpersonal Violence, 16, 361-
380.
Gottfredson, M. R., & Hirschi, T. (1990). A general theory of crime. Stan
ford, CA: Stanford University Press.
Grann, M., & Wedin, I. (in press). Risk factors for recidivism among
spousal assault and spousal homicide offenders. Psychology, Crime
& Law.
Gretton, H., McBride, M., & Hare, R. D. (1995). Psychopathy in adolescent
sex offenders: A follow-up study. Paper presented at the annual
conference of the Association for the Treatment of Sexual Abusers,
New Orleans, LA.
Gretton, H., McBride, M., O’Shaughnessy, R., & Hare, R. D. (1994). Pre
dicting patterns of criminal activity in adolescent sexual psychopaths
[Abstract]. Canadian Psychology, 35, 50.
Gretton, H., McBride, M., O’Shaughnessy, R., & Hare, R. D. (1997). Sex
offender or generalized offender? Psychopathy as a risk marker for
violence in adolescent offenders [Abstract]. Canadian Psychology,
38( 2a), 15.
Hanson, R. K. (1997). How to know what works with sex offenders. Sex
Abuse: A Journal of Research and Treatment, 9 ,129-145.
Hanson, R. K., & Bussiere, M. T. (1998). Predicting relapse: A meta
analysis of sexual offender recidivism studies. Journal of Consulting
and Clinical Psychology, 66, 348-362.
129
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Hanson, R. K., Cadsky, O., Harris, A., & Lalonde, C. (1997). Correlates of
battering among 997 men: Family history, adjustment, and attitudi-
nal differences. Violence and Victims, 12,191-208.
Hanson, R. K., & Harris, A. J. R. (2000). Where should we intervene? Dy
namic predictors of sexual offense recidivism. Criminal Justice and
Behavior, 27, 6-35.
Hanson, R. K., Scott, H., & Steffy, R. A. (1995). A comparison of child
molesters and non-sexual criminals: Risk predictors and long-term
recidivism. Journal of Research in Crime and Delinquincy, 32, 325-
327.
Hare, R. D. (1991). Manual for the Hare Psychopathy Checklist-Revised.
Toronto, Canada: Multi-Health Systems.
Hare, R. D. (1996). Psychopathy: A clinical construct whose time has
come. Criminal Justice and Behavior, 23, 25-54.
Hare, R. D. (1998). The Hare PCL-R: Some issues concerning its use and
misuse. Legal and Criminological Psychology, 3, 99-119.
Hare, R. D., Harpur, T. J., Hakstian, A. R., Forth, A. E., Hart, S. D., & New
man, J. P. (1990). The Revised Psychopathy Checklist: Descriptive
statistics, reliability, and factor structure. Psychological Assessment,
2, 338-341.
Harpur, T. J., Hakstian, A. R., & Hare, R. D. (1988). Factor structure of the
Psychopathy Checklist. Journal of Consulting and Clinical Psychol
ogy, 56, 741-747.
Harpur, T. J., Hare, R. D., & Hakstian, A. R. (1989). Two-factor conceptu
alization of psychopathy: Construct validity and assessment implica
tions. Psychological Assessment: A Journal of Consulting and
Clinical Psychology, 1, 6-17.
Harris, G. T., & Rice, M. E. (1997). Risk appraisal and management of
violent behavior. Psychiatric Services, 48, 1168-1176.
Harris, G. T., Rice, M. E., & Cormier, C. A. (1991). Psychopathy and vio
lent recidivism. Law and Human Behavior, 15, 625-637.
Harris, G. T., Rice, M. E., & Quinsey, V. L. (1993). Violent recidivism of
mentally disordered offenders: The development of a statistical pre
diction instrument. Criminal Justice and Behavior, 20, 315-335.
Hart, S. D., Cox, D., & Hare, R. D. (1995). Manual for the Screening Ver
sion of the Hare Psychopathy Checklist-Revised (PCL:SV).
Toronto: Multi-Health Systems.
130
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Hart, S. D., & Hare, R. D. (1996). Psychopathy and risk assessment. Cur
rent Opinion in Psychiatry, 9, 380-383.
Hart, S. D., Hare, R. D., & Forth, A. E. (1994). Psychopathy as a risk
marker for violence: Development and validation of a screening ver
sion of the Revised Psychopathy Checklist. In J. Monahan & H. J.
Steadman (Eds.), Violence and mental disorder: Developments in
risk assessment (pp. 81 -97). Chicago: University of Chicago Press.
Hart, S. D., Kropp, P. R., & Hare, R. D. (1988). Performance of male psy
chopaths following conditional release from prison. Journal of Con
sulting and Clinical Psychology, 56, 227-232.
Hathaway, S. R., & McKinley, J. C. (1983). The Minnesota Multiphasic
Personality Inventory manual. New York: Psychological Corpora
tion.
Heilbrun, K., Hart, S. D., Hare, R. D., Gustafson, D., Nunez, C., & White, A.
(1998). Inpatient and post-discharge aggression in mentally disor
dered offenders: The role of psychopathy. Journal of Interpersonal
Violence, 13, 514-527.
Hemphill, J. F., Hare, R. D., & Wong, S. (1998). Psychopathy and recidi
vism: A review. Legal and Criminological Psychology, 3 ,139-170.
Hemphill, J. F., Templeman, R., Wong, S., & Hare, R. D. (1998). Psycho
pathy and crime: Recidivism and criminal careers. In D. J. Cooke,
A. E. Forth, & R. D. Hare (Eds.), Psychopathy: Theory, research,
and implications for society (pp. 375-399). Dordrecht, The Nether
lands: Kluwer.
Hiday, V. A., Swartz, M. S., Swanson, J. W., Borom, R., & Wagner, H. R.
(1998). Male-female differences in the setting and construction of
violence among people with severe mental illness. Social Psychiatry
and Psychiatric Epidemiology, 33(Supp. 1), S68-S74.
Hill, C. D. (1997). Prediction of aggressive and socially disruptive behavior
among forensic patients: A validation study of the Psychopathy
Checklist-Screening Version. Dissertation Abstracts International, #.
Hill, C. D., Rogers, R., & Bickford, M. E. (1996). Prediction aggressive and
socially disruptive behavior in a maximum security forensic hospital.
Journal of Forensic Sciences, 41, 56-59.
Hillbrand, M. (1995). Aggression against self and aggression against oth
ers in violent psychiatric patients. Journal of Consulting and Clinical
Psychology, 63, 668-671.
131
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Hilton, N. Z., Harris, G. T., & Rice, M. E. (2001). Predicting violence by
serious wife assaulters. Journal of Interpersonal Violence, 16, 408-
423.
Hobson, J., & Shine, J. (1998). Measurement of psychopathy in a UK
prison population referred for long-term psychotherapy. British Jour
nal of Criminology, 38, 504-515.
Hodgins, S. (1990). Prevalence of mental disorders among penitentiary
inmates in Quebec. Canada’ s Mental Health, 3 7 ,1 -4.
Hodgins, S., Mednick, S. A., Brennan, P. A., Schulsinger, F., & Engberg, M.
(1996). Mental disorder and crime: Evidence from a Danish birth
cohort. Archives of General Psychiatry, 53, 489-496.
Hoffman, P. B., & Beck J. L. (1985). Recidivism among released federal
prisoners: Salient factor score and five-year follow-up. Criminal
Justice and Behavior, 12, 501-507.
Hogarty, G. E., Goldberg, S. C., & Schooler, N. (1974). Drug and socio
therapy in the aftercare of schizophrenic patients: Adjustments of
non-relapsed patients. Archives of General Psychiatry, 31, 340-347.
Hwu, H. G., Yeh, E. K., & Change, L. Y. (1989). Prevalence of psychiatric
disorders in Taiwan defined by the Chinese Diagnostic Interview
Schedule. Acta Psychiatrica Scandinavica, 7 9 ,136-147.
Kaplan, H. I., & Sadock, B. J. (1996). Concise textbook of clinical psychia
try. Baltimore: Williams and Wilkins.
Khantzian, E. J. (1997). The self-medication hypothesis of substance use
disorders: A reconsideration and recent applications. Harvard Re
view of Psychiatry, 4, 231-244.
Klassen, D., & O’Connor, W. A. (1994). Demographic and case history
variables in risk assessment. In J. Monahan & H. J. Steadman
(Eds.), Violence and mental disorder: Developments in risk assess
ment (pp. 229-258). Chicago: University of Chicago Press.
Kopelowicz, A., Liberman, R. P., & Zarate, R. (2002). Psychosocial treat
ments for schizophrenia. In P. E. Nathan & J. M. Norman (Eds.),
A guide to treatments that work (2nd ed., pp. 201-228). London:
Oxford University Press.
Koskinen, O., Sauvola, A., Valonen, P., Hakko, H., Marjo-Riitta, J., & Rae-
saenen, P. (2001). Increased risk of violent recidivism among adult
males is related to singe-parent family during childhood: The North
ern Finland 1966 Birth Cohort Study. Journal of Forensic Psychiatry,
12, 539-548.
132
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Kosson, D. S., Smith, S. S., & Newman, J. P. (1990). Evaluating the con
struct validity of psychopathy in Black and White inmates: Three
preliminary studies. Journal of Abnormal Psychology, 99, 250-259.
Langan, P. A., & Levin, D. J. (2002). Bureau of Justice Statistics special
report: Recidivism of prisoners released in 1994. Washington, DC:
U.S. Department of Justice, Office of Justice Programs.
Lidz, C. W., Mulvey, E. P., & Gardner, W. (1993). The accuracy of predic
tions of violence to others. Journal of the American Medical Asso
ciation, 2 6 9 ,1007-1111.
Link, B. G., & Stueve, A. (1994). Psychotic symptoms and the violent/ille
gal behavior of mental patients compared to community controls. In
J. Monahan & H. J. Steadman (Eds.), Violence and mental disorder:
Developments in risk assessment (pp. 137-159). Chicago: Univer
sity of Chicago Press.
Lipsey, M. W. (1992). The effect of treatment on juvenile delinquents: Re
sults from meta-analysis. In F. Losel, D. Bender, & T. Blieser (Eds.),
Psychology and law: International perspectives (pp. 131-143). New
York: de Gruyter.
Lipsey, M. W. (1995). What do we learn from 400 research studies on the
effectiveness of treatment with juvenile delinquents? In J. McGuire
(Ed.), What works: Reducing reoffending, guidelines from research
and practice (p. 78). West Sussex, England: Wiley.
Losel, F. (1995). The efficacy of correctional treatment: A review and
synthesis of meta-evaluations. In J. McGuire (Ed.), What works:
Reducing reoffending, guidelines from research and practice (pp.
79-111). West Sussex, England: Wiley.
Loza, W. (1993a). Different substance abusing offenders require a unique
program. International Journal of Offender Therapy and Compara
tive Criminology, 37, 351-358.
Loza, W. (1993b). Effectiveness of the Wisconsin Case Management
strategies for use with Canadian offenders. Dissertation Abstracts
International, 54(2-B), 1139.
MacCulloch, M. J., Snowden, P. R., & Wood, P. J., & Mills, M. (1983).
Sadistic fantasy, sadistic behavior and offending. British Journal of
Psychiatry, 143, 20-29.
Maletzky, B. M. (1993). Factors associated with success and failure in the
behavioral and cognitive treatment of sexual offenders. Annals of
Sex Research, 6, 241-258.
133
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Malm, U. (1982). The influence of group therapy on schizophrenia. Acta
Psychiatrica Scandanavica, 65(Suppl.), 65-73.
Marder, S. R., Wirshing, W. C., Mintz, J., McKenzie, J., Johnston-Cronk, K.,
Eckman, T. A., etal. (1996). Two-year outcome of social skills
training and group psychotherapy for outpatients with schizophrenia.
American Journal of Psychiatry, 153,1585-1592.
McNiel, D. (1994). Hallucinations and violence. In J. Monahan & H.
Steadman (Eds.), Violence and mental disorder: Developments in
risk assessment (pp. 83-202). Chicago: University of Chicago Press.
Mednick, S. A. (1977). A biosocial theory of the learning of law-abiding
behavior. In S. A. Mednick & K. O. Christiansen (Eds.), Biosocial
bases of criminal behavior (pp. 1-8). New York: Gardner.
Mednick, S. A., & Kandel, E. (1988). Genetic and perinatal factors in vio
lence. In S. A. Moffitt & T. Moffitt (Eds.), Biological contributions to
crime causation (pp. 121-134). Dordrecht, Holland: Martinus Nijhoff.
Millon, T., Simonsen, E., & Birket-Smith, M. (1998). Historical conceptions
of psychopathy in the United States and Europe. In T. Millon, E.
Simonsen, M. Birket-Smith, & R. Davis (Eds.), Psychopathy: Antiso
cial, criminal, and violent behavior (Chap. 1). New York: Guilford.
Moffitt, T. E. (1993). Adolescence-limited and life-course-persistent antiso
cial behavior: A developmental taxonomy. Psychological Review,
100, 674-701.
Monahan, J. (1992). Mental disorder and violent behavior: Perceptions
and evidence. American Psychologist, 47 ,511-521.
Monahan, J., & Steadman, H. (Eds.). (1994). Violence and mental disor
der: Developments in risk assessment. Chicago: University of Chi
cago Press.
Monahan, J., Steadman, H. J., Silver, E., Appelbaum, P. S., Robbins, P. C.,
Mulvey, E. P., et al. (2001). Rethinking risk assessment: The
MacArthur study of mental disorder and violence. New York: Oxford
Usniversity Press.
Mossman, D. (1994). Further comments on portraying the accuracy of
violence predictions. Law and Human Behavior, 18, 587-593.
Mulvey, E. P. (1994). Assessing the evidence of a link between mental
illness and violence. Hospital and Community Psychiatry, 45, 663-
668.
Mulvey, E. P., & Lidz, C. W. (1993). Measuring patient violence in danger
ousness research. Law and Human Behavior, 17, 277-288.
134
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Nolan, K. A., Volavka, J., Mohr, P, & Czobor, P. (1999). Psychopathy and
violent behavior among patients with schizophrenia or schizoaffec
tive disorder. Psychiatric Services, 50, 787-792.
Novaco, R. W. (1994). Anger as a risk factor for violence among the
mentally disordered. In J. Monahan & H. J. Steadman (Eds.), Vio
lence and mental disorder: Developments in risk assessment (pp.
21-60). Chicago: University of Chicago Press.
Ogloff, J. R. P., Wong, S., & Greenwood, A. (1990). Treating criminal psy
chopaths in a therapeutic community program. Behavioral Sciences
and the Law, 8, 181-190.
Overall, J. E., & Klett, C. J. (1962). The Brief Psychiatric Rating Scale.
Psychological Reports, 10, 799-812.
Paris, J. (1998). A biopsychosocial model of psychopathy. In T. Millon, E.
Simonsen, M. Birket-Smith, & R. D. Davis (Eds.), Psychopathy: An
tisocial, criminal, and violent behavior {pp. 277-287). New York:
Guilford.
Parker, R. N., & Auerhahn, K. (1998). Alcohol, drugs, and violence.
Annual Review of Sociology, 24, 291 -311.
Patterson, G. R., & Yoerger, K. (1993). Developmental models for delin
quent behavior. In S. Hodgins (Ed.), Mental disorder and crime (pp.
140-172). Thousand Oaks, CA: Sage.
Pinard, L., & Pagani, G.-F. (2001). Clinical assessment of dangerousness.
Cambridge: Cambridge University Press.
Pinel, P. (1962). A treatise on insanity (D. Davis, Trans.). New York: Haf-
ner. (Original work published 1801)
Polloch, V. E., Briere, J., Schneider, L., Knop, J., Mednick, S. A., & Good
win, D. W. (1990). Childhood antecedents of antisocial behavior:
Parental alcoholism and physical abusiveness. American Journal of
Psychiatry, 147, 1290-1293.
Porter, S., Birt, A. R., & Boer, D. P. (2001). Investigation of the criminal
and conditional release profiles of Canadian federal offenders as a
function of psychopathy and age. Law and Human Behavior, 25,
647-661.
Quinsey, V. L., Harris, G. T., Rice, M. E., & Cormier, C. A. (1998). Violent
offenders: Appraising and managing risk. Washington, DC: Ameri
can Psychological Association.
Quinsey, V. L., Rice, M. E., & Harris, G. T. (1995). Actuarial prediction of
sexual recidivism. Journal of Interpersonal Violence, 10, 85-105.
135
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Rabinowitz, J., & Mark, M. (1999). Risk factors for violence among long-
stay psychiatric patients: National study. Acta Psychiatrics Scandi-
navica, 99, 341-347.
Raesaenen, P., Tiihonen, J., Isohanni, M., Rantakallio, P., Lehtonen, J., &
Moring, J. (1998). Schizophrenia, alcohol abuse, and violent be
havior: A 26-year follow-up study of an unselected birth cohort.
Schizophrenia Bulletin, 24, 437-441.
Raine, A. (1993). The psychopathology of crime: Criminal behavior as a
clinical disorder. San Diego: Academic Press.
Raine, A. (2002). Biosocial studies of antisocial and violent behavior in
children and adults: A review. Journal of Abnormal Child Psychol
ogy, 30, 311 -326.
Raine, A., Brennan, P. A., & Mednick, S. A. (1994). Birth complications
combined with early maternal rejection at age 1 year predispose to
violent crime at age 18 years. Archives of General Psychiatry, 51,
984-988.
Rasmussen, K., & Levander, S. (1995). Violence in the mentally disor
dered: A differential clinical perspective. Issues in Criminological
and Legal Psychology, 2 4 ,127-130.
Rice, M. E., & Harris, G. T. (1992). A comparison of criminal recidivism
among schizophrenic and nonschizophrenic offenders. International
Journal of Law and Psychiatry, 15, 397-408.
Rice, M. E., & Harris, G. T. (1995). Psychopathy, schizophrenia, alcohol
abuse, and violent recidivism. International Journal of Law and Psy
chiatry, 18, 333-342.
Rice, M. E., & Harris, G. T. (1997). Cross-validation and extension of the
Violence Risk Appraisal Guide for child molesters and rapists. Law
and Human Behavior, 21, 231 -241.
Rice, M. E., Harris, G. T., & Cormier, C. A. (1992). An evaluation of a
maximum security therapeutic community for psychopaths and other
mentally disordered offenders. Law and Human Behavior, 16, 399-
412.
Rice, M. E., Harris, G. T., & Quinsey, V. L. (1990). A follow-up of rapists
assessed in a maximum security psychiatric facility. Journal of Inter
personal Violence, 5, 435-448.
Roberts, T. K., Mock, L A. T., & Johnstone, E. E. (1981). Psychological
aspects of the etiology of violence. In J. R. Hays, T. K. Roberts, &
K. S. Solway (Eds.), Violence and the violent individual (pp. 9-33).
New York: Spectrum.
136
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Robins, L. N. (1966). Deviant children grown up: A sociologoical and psy
chiatric study of sociopathic personality. Oxford, England: Williams
and wilkins.
Robins, L. N., & Regier, D. A. (Eds.). (1991). Psychiatric disorders in
America: The Epidemiologic Catchment Area study. New York:
Free Press.
Rogers, R. (1995). Diagnostic and structured interviewing: A handbook for
psychologists. Odessa, FL: Psychological Assessment Resources.
Ross, D. J., Hodgins, S., & Cote, G. (1992). The predictive validity of the
French Psychopathy Checklist: Male inmates on parole. Montreal,
Quebec: University of Montreal, Department of Psychology.
Rowe, D. C. (2001). Biology and crime. Los Angeles: Roxbury.
Rutter, M. L. (1997). Nature-nurture integration: The example of antisocial
behavior. American Psychologist, 52, 390-398.
Rutter, M., & Madge, N. (1976). Cycles of disadvantage: A review of re
search. London: Heineman.
Salekin, R. T., Rogers, R., & Sewell, K. W. (1996). A review and meta
analysis of the Psychopathy Checklist and Psychopathy Checklist-
Revised: Predictive validity of dangerousness. Clinical Psychology:
Science and Practice, 3, 203-215.
Saunders, D. G. (1995). Prediction of wife assault. In J. C. Campbell
(Ed.), Assessing dangerousness: Violence by sexual offenders,
batterers, and child abusers (pp. 68-95). Thousand Oaks, CA:
Sage.
Schulsinger, F. (1972). Psychopathy: Heredity and environment. Interna
tional Journal of Mental Health, 1 ,190-206.
Scott, H., Johnson, S., Menezes, P., Thornicroft, G., Marshall, J., Bindman,
J.,etal. (1998). Substance misuse and risk of aggression and of
fending among the severely mentally ill. British Journal of Psychia
try, 172, 345-350.
Serin, R., & Lawson, J. S. (1987). Prediction of temporary absence out
come for penitentiary inmates. Canadian Journal of Criminology, 29,
35-49.
Serin, R. C. (1991). Psychopathy and violence in criminals. Journal of
Interpersonal Violence, 6, 423-431.
Serin, R. C. (1996). Violent recidivism in criminal psychopaths. Law and
Human Behavior, 20, 207-217.
137
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Serin, R. C., & Amos, N. L. (1995). The role of psychopathy in the assess
ment of dangerousness. International Journal of Law and Psychia
try, 18, 231-238.
Serin, R. C., Mailloux, D. L., & Malcolm, P. B. (2001). Psychopathy, devi
ant sexual arousal, and recidivism among sexual offenders. Journal
of Interpersonal Violence, 16, 234-346.
Serin, R. C., Peters, R. D., & Barbaree, H. E. (1990). Predictors of psycho
pathy and release outcome in a criminal population. Psychological
Assessment, 2, 419-422.
Seto, M. C., & Barbaree, H. E. (1999). Psychopathy, treatment behavior,
and sex offenders recidivism. Journal of Interpersonal Violence, 14s,
1235-1248.
Sheitman, B. B., Kinon, B. J., Ridgway, B. A., & Lieberman,J. (1998).
Pharmacological treatments of schizophrenia. In P. Nathan & J.
Gorman (Eds.), A guide to treatments that work (pp. 167-189). Lon
don: Oxford University Press.
Shrout, P., Lyons, M., Dohrenwend, B. P., Skodol, A., Solomon, M., & Kass,
F. (1988). Changing time frames on symptom inventories: Effects
on the Psychiatric Epidemiology Research Interview. Journal of
Consulting and Clinical Psychology 56, 267-272.
Skeem, J., & Mulvey, E. (2001a). Assessing the violence potential of
mentally disordered offenders being treated in the community. In A
Buchanan (Ed.), Care of the mentally disordered offender in the
community (pp. 104-121). Oxford: Oxford University Press.
Skeem, J., & Mulvey, E. (2001b). Psychopathy and community violence
among civil psychiatric patients: Results from the MacArthur Vio
lence Risk Assessment Study. Journal of Consulting and Clinical
Psychology, 69, 358-374.
Snyder, K. S., Wallace, C. J., Moe, K., & Liberman, R. P. (1994). Ex
pressed emotion by residential care operators and schizophrenic
residents’ symptoms and quality of life. Hospital and Community
Psychiatry, 4 5 ,1141 -1143.
Sreenivasan, S., Kirkish, P., & Eth, S. (1997). Predictors of recidivistic
violence in criminally insane and civilly committed psychiatric inpa
tients. International Journal of Law and Psychiatry, 20, 279-291.
Steadman, H., Monahan, J., Robbins, P., Appelbaum, P., Grisso, T., Klas-
sen, D., et al. (1993). From dangerousness to risk assessment: Im
plications for appropriate research stragies. In S. Hodgins (Ed.),
Mental disorder and crime (pp. 39-62). Newbury Park, CA: Sage.
138
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Steadman, H. J., & Cocozza, J. J. (1974). Careers of the criminally insane:
Excessive social control of deviance. Lexington, MA: Lexington
Books.
Steadman, H. J., Monahan, J., Appelbaum, P. S., Grisso, T., Mulvey, E. P.,
Roth, L. H., et al. (1994). In J. Monahan & H. J. Steadman (Eds.),
Violence and mental disorder (pp. 297-318). Chicago: University of
Chicago Press.
Steadman, H. J., Mulvey, E., Monahan, J., Robins, P. C., Appelbaum, P. I.
S., Grisso, T., et al. (1998). Violence by people discharged from
acute psychiatric inpatient facilities and by others in the same
neighborhoods. Archives of General Psychiatry, 55, 393-401.
Swanson, J. W. (1994). Mental disorder, substance abuse, and community
violence: An epidemiological approach. In J. Monahan & H. J.
Steadman (Eds.), Violence and mental disorder: Developments in
risk assessment (pp. 101-136). Chicago: University of Chicago
Press.
Swanson, J. W., Borum, R., Swartz, M. S., & Monahan, J. (1996). Psy
chotic symptoms and disorders and the risk of violent behavior in the
community. Criminal Behavior and Mental Health, 6, 317-338.
Swanson, J. W., Holzer, C. E., Ganju, V. K., & Jono, R. T. (1990). Vio
lence and psychiatric disorder in the community: Evidence from the
Epidemiologic Catchment Area surveys. Hospital and Community
Psychiatry, 41, 761-770.
Swartz, M. S., Swanson, J. W., Hiday, V. A., Borum, R., Wagner, H. R., &
Burns, B. J. (1998). Violence and severe mental illness: The ef
fects of substance abuse and nonadherence to medication. Ameri
can Journal of Psychiatry, 155, 226-231.
Tardiff, K., Marzuk, P. M., Leon, A. C., Portera, L., & Weiner, C. (1997).
Violence by patients admitted to a private psychiatric hospital.
American Journal of Psychiatry, 154, 88-93.
Taylor, P. J. (1985). Motives for offending among violent and psychotic
men. British Journal of Psychiatry, 147, 491-498.
Tedeschi, J. T., & Felson, R. B. (1994). Violence, aggression, and coer
cive action. Washington, DC: American Psychological Association.
Tengstrom, A., Grann, M., L&ngstrom, N., & Kullgren, G. (2000). Psycho
pathy (PCL-R) as a predictor of violent recidivism among criminal
offenders with schizophrenia. Law and Human Behavior, 24{ \ ), 45-
58.
139
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Thornberry, T. P., & Jacoby, J. E. (1979). The criminally insane: A com
munity follow-up of mentally ill offenders. Chicago: University of
Chicago Press.
Tinsley, E. A., & Weiss, David J. (1975). Interrater reliability and agree
ment of subjective judgments. Journal of Counseling Psychology,
22, 358-376.
Toupin, J., Mercier, H., Dery, M., Cote, G., & Hodgins, S. (1996). Validity
of the PCL-R for adolescents. In D. J. Cooke, A. E. Forth, J. P.
Newman, & R. D. Hare (Eds.), Issues in criminological and legal
psychology: International perspectives on psychopathy (pp. 143-
145). Leicester, UK: British Psychological Society.
Turner, W. M., & Tsuang, M. T. (1990). The impact of substance abuse on
the course and outcome of schizophrenia. Schizophrenia Bulletin,
16, 87-95.
Valliant, P. M., Gristey, C., Pottier, D., & Kosmyna, R. (1999). Risk factors
in violent and nonviolent offenders. Psychological Reports, 85, 675-
680.
Vida, S. (1993). A computer program for non-parametric receiver operat
ing characteristic analysis. Computer Methods and Programs in
Biomedicine, 40, 95-101.
Vida, S. (2001). AccuROC for Windows 95/98/NT Version 2.4, May, 2001.
Montreal, Quebec, Canada: McGill University Health Centre. Avail
able at http://www.Accumetric.com (manual was part of software
program purchased and downloaded from this Web site)
Wallace, J. (1989). A biopsychosocial model of alcoholism. Social Case
work: The Journal of Contemporary Social Work, 7, 325-332.
Webster, C. D., Douglas, K. S., Eaves, D., & Hart, S. D. (1997). Predicting
violence in mentally and personality disordered individuals. In C. D.
Webster & M. A. Jackson (Eds.), Impulsivity: Theory, assessment,
and treatment (pp. 135-177). New York: Guildford.
Webster, C. D., Eaves, D., Douglas, K. S., & Wintrup, A. (1995). The
HCR-20 scheme: The assessment of dangerousness and risk.
Vancouver, Canada: Simon Fraser University and British Columbia
Forensic Psychiatric Services Commission.
Weiss, G., & Hechtman, L. T. (1993). Hyperactive children grown up:
ADHD in children, adolescents, and adults (2nd ed.). New York:
Guilford.
Wessely, S. (1998). The Camberwell study of crime and schizophrenia.
Social Psychiatry and Psychiatric Epidemiology, 33, S24-S28.
140
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
West, D. J., & Farrington, D. P. (1973). Who becomes delinquent? Sec
ond report of the Cambridge Study in Delinquent Development.
Oxford, England: Crane, Russak.
White, R. J., Ackerman, R. J., & Caraveo, L. E. (2001). Self-identified
alcohol abusers in a low-security federal prison: Characteristics and
treatment implications. International Journal of Offender Therapy
and Comparative Criminology, 45, 214-227.
Widiger, T., & Trull, T. (1994). Personality disorders and violence. In J.
Monahan & H. J. Steadman (Eds.), Violence and mental disorder
(pp. 203-226). Chicago: University of Chicago Press.
Wiederanders, M., & Choate, P. A. (1994). Beyond recidivism: Measuring
community adjustments of conditionally released insanity acquittees.
Psychological Assessment, 6, 61-66.
Wintrup, A. (1996). Assessing risk of violent in mentally disordered of
fenders with the HCR-20. Unpublished master’s thesis, Simon
Fraser University, Burnaby, British Columbia.
Wintrup, A., Coles, M., Hart, S. D., & Webster, C. D. (1994). The predictive
validity of the PCL-R in high-risk mentally disordered offenders [Ab
stract]. Canadian Psychology, 35(2a), 47.
Wyatt, R. J., Kirch, D. G., & Egan, M. F. (1995). Schizophrenia: Neuro
chemical, viral, and immunological studies. In H. I. Kaplan & B. J.
Sadock (Eds.), Comprehensive textbook of psychiatry (6th ed., p.
95). Baltimore: Williams and Wilkins.
Yesavage, J. A., Benezech, M., Larrieu-Arguille, R., & Bourgeois, M.
(1986). Recidivism of the criminally insane in France: A 22-year
follow-up. Journal of Clinical Psychiatry, 47, 465-466.
Zamble, E., & Quinsey, V. L. (1997). The criminal recidivism process.
Cambridge, England: Cambridge University Press.
Zamble, E., & Palmer, W. (1996). Prediction of recidivism using psycho
pathy and other psychologically meaningful variables. In D. J.
Cooke, A. E. Forth, J. P. Newman, & R. D. Hare (Eds.), Issues in
criminological and legal psychology: International perspectives on
psychopathy (pp. 153-156). Leicester, UK: British Psychological
Society.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Predictors of dysphoria and substance abuse among Latino and African American youth
PDF
Psychologists' perceptions of older clients: The effect of age, gender, knowledge, and experience
PDF
Parenting as a moderator of minority adolescent victimization and violent behavior in high -risk neighborhoods
PDF
Early adolescent drug use among multiethnic males: A prospective examination of the influences of psychological distress, relationship with family and school, law abidance, guilt and peer drug use
PDF
Using the BASC Parent Rating and Self -Report composite measures to differentiate juvenile sex offenders from delinquent non -sex offending adolescents
PDF
Mental and physical health consequences of intimate partner violence in a multi-ethnic sample of women
PDF
The therapy relationship: Developing a sound metaphysical description of personal relations
PDF
Impact of language and culture on a neuropsychological screening battery for Hispanics
PDF
The genetic link between violence and alcoholism
PDF
Self -regulated learning of medical students: Assessment of a social cognitive model
PDF
The roles of acculturation and family relationships in the career decision self -efficacy of Asian American college students
PDF
The role of sensation seeking in high risk sexual behavior of adult male inmates from varying family structures
PDF
Protective factors influencing resiliency among African -American and Latino students
PDF
The relationship among attachment style, self-esteem, and violence among incarcerated male inmates
PDF
Predictors and consequences of drug abuse in adulthood: Associations with psychological distress, intimacy, work adjustment and criminal behavior
PDF
Treating the new untouchables: Mental health professionals and sex offender treatment (a short radio documentary)
PDF
Psychological well-being, religious affiliation, and cultural identification in adult children of Jewish-Gentile interfaith marriages
PDF
Substance use and disordered eating as outcomes of childhood maltreatment in a community sample of adolescent Latina females
PDF
Marital relationships in Taiwanese international students: A qualitative approach
PDF
Stress, appraisal and coping among Japanese-American, Anglo-American, African-American and Mexican-American spousal caregivers of persons with dementia
Asset Metadata
Creator
Blum, Frederick M.
(author)
Core Title
Psychopathy, psychosis, drug abuse, and reoffense among conditionally released offenders
Degree
Doctor of Philosophy
Degree Program
Education (Counseling Psychology)
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
health sciences, mental health,OAI-PMH Harvest,sociology, criminology and penology
Language
English
Contributor
Digitized by ProQuest
(provenance)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-632445
Unique identifier
UC11340119
Identifier
3116668.pdf (filename),usctheses-c16-632445 (legacy record id)
Legacy Identifier
3116668.pdf
Dmrecord
632445
Document Type
Dissertation
Rights
Blum, Frederick M.
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
health sciences, mental health
sociology, criminology and penology