University of Southern California Dissertations and Theses

Typologies of male and female adolescent perpetrators of dating violence: three subtypes for males and four subtypes for females, and differences among them

(USC Thesis Other)

Typologies of male and female adolescent perpetrators of dating violence: three subtypes for males and four subtypes for females, and differences among them

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

TYPOLOGIES OF MALE AND FEMALE ADOLESCENT PERPETRATORS OF
DATING VIOLENCE:
THREE SUBTYPES FOR MALES AND FOUR SUBTYPES FOR FEMALES,
AND DIFFERENCES AMONG THEM

by

Christopher Michael Thomson

__________________________________________________________________

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
(SOCIAL WORK)

May 2008

Copyright 2008 Christopher Michael Thomson

ii
Acknowledgments
There are so many people I need to thank for the reassurance, support,
guidance, advocacy, wisdom, spirit, genes, and love. Therefore, I need to start at the
beginning. I need to thank my parents, Larry and Carol Thomson, for raising me and
instilling the tenacity and grit necessary to continue this project at all costs. I also
would like to thank my sister Carrie Allen and her son Chase Allen for their
unwavering belief in me. There were several other people who have supported me
along the way, such as Dr. Janet Eddy and USC staff at the Learning Center, Dr.
Elise Brisco, and Dr. Bill McClure. I am very grateful for your guidance and support
along the way. I also want to thank my best friend and colleague, Jason Sackett,
LCSW. You were truly amazing. I am very grateful to your editing comments and
feedback on the plethora of manuscripts I had you read. I always knew we were
close to completing a section of the dissertation when you would say, “I think I
understand it.” I love you, man. Also, many thanks to Jason’s wife Karen Sackett,
for allowing Jason the time to proof read my many texts when I am sure there were a
thousand other things she would rather he being doing.
Speaking of amazing wives, it is time for me to mention and thank a truly
extraordinary person, wife, mother, and saint, my wife, Toshi Kuramatsu, LCSW.
You are my love. This dissertation is a testament of our love for one another. We did
it, babe! I could have never done it without you. You are my rock. You never
faltered. I could always count on you. I love you very much. Thanks for everything.
I would also like to thank my daughter, Manishaa Kuramatsu –Thomson, who was

iii
my inspiration. When times were tough and I didn’t think I was going to make it, I
thought of you, Manishaa. You gave me strength when I didn’t think I had any left.
You gave me hope when things looked bleak, comforted me when I was
discouraged, and made me laugh when I was sad. I love you, sweetheart. You are
the best.
Next, I would like to say a special thanks to Dr. John Brekke and Dr. Ramon
Salcido who both provided mentorship and guidance that led to me embarking on my
journey to successfully complete a Ph.D. in Social Work at USC. I am very grateful
to both of you. I would also like to thank Dr. Robert Nishimoto for his teaching,
insight, and friendship. Dr. Maura O’Keefe is a very special person that deserves
much of the credit in terms of helping me bring this dissertation to a successful
conclusion. Maura, I cannot thank you enough for hanging in there with me. You
were fantastic. Thank you so much for your guidance, support, encouragement,
generosity, and kindness. I will always be grateful for your assistance. I would like
to make a special thanks to Dr. Robert Moore. I am so very grateful for your
presence, support, and guidance. I have learned so much over the years from you as
my Zen teacher. I feel so fortunate to have had you on my committee. It was truly a
blessing. Dr. Brad Zebrack, you are truly a gift from God. I am so very appreciative
of your willingness to serve on my committee both as a member and later as my
chair. Your clarity and guidance was essential in helping me to complete this
process. I am very grateful. I would like to thank Malinda Sampson, Dr. Bruce
Jansson, and Dr. Marilyn Flynn for their assistance and support. I would also like to

iv
thank Dr. Shantanu DuttaAhmed for his excellent feedback and guidance in helping
me to complete the final edit of my dissertation.
There are a few more very special people that helped make my work
possible. First, I would like to thank three of my clinical colleagues, Dr. Greg
Wheeler, Dr. Keith Valone, and Dr. Bonnie Saland. I cannot thank each of you
enough for your support, kind words, and encouragement. You will always be
special people in my life. Second, I would like to thank Dr. David Wexler and Dr.
Gwen Uman. Both of you provided excellent consultation related to this project. I
am very grateful for Dr. Wexler’s feedback on earlier drafts of my manuscript,
particularly the literature review. Your assistance definitely pointed me in the right
direction, and I want you to know how valuable that was to me. I am also very
grateful to Dr. Uman for going far above the call of duty on this project. Your
compassionate coaching style for the statistical and research methods portions of my
dissertation was absolutely wonderful, not to mention the umpteen manuscripts you
reviewed to help improve those sections. Thanks for everything. You are fantastic.
Third, I would like to thank my friend and attorney, Godfrey Isaac, a remarkable
man. I am overwhelmed by your generosity, wisdom, and kindness. Thank you so
very much for all your support. Fourth, I would like to thank Dr. Julena Lind. It is
difficult to find the words to thank you. Every time I think of what you did for me, I
get a little emotional. I came to you with a dream and you set me free to make that
dream a reality. It was truly an amazing feat! You are remarkable, a true gift. Thank
you again.

v
My fifth and final thanks go to my clients over the last twenty years of
working in the mental health field. You are my greatest teachers. You have taught
me so much. I could not have done this without you. You made me a better person,
student, teacher, researcher, and therapist. I am truly humbled by each and every one
of you. Thank you all so very much. This is a miracle that has come true and I am
truly grateful to everyone who has helped me. You are all very special angels in my
life.

vi
Table of Contents
Acknowledgments ……………………………………………………….. ii
List of Tables …………………………………………………………….. vii
List of Figures …………………………………………………………….. ix
Abstract ……………………………………………………………………. x
Chapter 1: Introduction …………………………………………………… 1
Chapter 2: Foundations of Typologies in Adolescent Perpetrators of
Dating Violence …………………………………………………………… 3

Chapter 3: Theoretical foundations and research objectives …………….. 25
Chapter 4: Methodology …………………………………………………. 38
Chapter 5: Results ……………………………………………………….. 58
Chapter 6: Discussion …………………………………………………… 88
References ………………………………………………………………. 109
Appendix A ………………………………………………………………118

vii
List of Tables
Table 1. Gender, Age, and SES of Each Adolescent
Male Typology Group ………………………………………………….. 59

Table 2. Gender, Age, and SES of Each Adolescent
Female Typology Group ……………………………………………….. 60

Table 3. Average Distance Between Adolescent Male Clusters ……… 63

Table 4. Average Distance Between Adolescent Female Clusters …….. 63

Table 5. Initial Cluster Centers for Males ……………………………... 64

Table 6. Initial Cluster Centers for Females …………………………… 65

Table 7. Final Cluster Centers for Males …………………………….... 66

Table 8. Final Cluster Centers for Females …………………………… 67

Table 9. Mean Scores by Cluster for Males …………………………... 72

Table 10. Mean Scores by Cluster for Females ………………………. 73

Table 11. Zero-order Pearson Product Correlation for
Overall Population ……………………………………………………. 75

Table 12. Partial Correlation When Controlling Age and SES for
Overall Population …………………………………………………….. 76

Table 13. Partial Correlation When Controlling Age and SES for
Male Generally Violent Group ………………………………………... 77

Table 14. Partial Correlation When Controlling Age and SES for
Male Situational Violent Group ………………………………………. 77

Table 15. Partial Correlation When Controlling Age and SES for
Male Non-Violent Group ……………………………………………. 78

Table 16. Partial Correlations When Controlling Age and SES for
Female Depressed Attention Seeker Group ………………………….. 80

viii
Table 17. Partial Correlations When Controlling Age and SES for
Female Hostile Controller Group …………………………………………. 80

Table 18. Partial Correlations When Controlling Age and SES for
Female Undifferentiated Controller Group ……………………………….. 81

Table 19. Partial Correlations When Controlling Age and SES for
Female Non-violent Group ……………………………………………….. 81

Table 20. Descriptive Statistics of Distal Risk Factors and Adolescent
Male Perpetrator Typology ………………………………………………. 84

Table 21. Descriptive Statistics of Proximal Risk Factors and
Adolescent Male Perpetrator Typology ………………………………….. 84

Table 22. Descriptive Statistics of Distal Risk Factors and
Adolescent Female Perpetrator Typology ……………………………….. 85

Table 23. Descriptive Statistics of Proximal Risk Factors and
Adolescent Female Perpetrator Typology ……………………………….. 85

Table 24. Male Adolescent Perpetrator Typologies: Distal and Proximal
Risk Factors of Dating Violence, Controlling for Age and SES ………… 86

Table 25. Female Adolescent Perpetrator Typologies: Distal and
Proximal Risk Factors of Dating Violence, Controlling for Age and SES ..87

ix
List of Figures
Figure 1. Riggs and O’Leary’s (1989) Dating Violence Model ……………… 15

Figure 2. Hypothesized Model: Typologies of male adolescent
perpetrators of dating violence and differences among them …………………. 39

Figure 3. Hypothesized Model: Typologies of female adolescent
perpetrators of dating violence and differences among them …………………. 39

x
Abstract
The study’s purpose was twofold: (i) identify subtypes for adolescent male
and female perpetrators of dating violence, and (ii) examine distal and proximal risk
factors of adolescent dating violence for differing male and female typologies. A
cluster analysis was used to develop perpetrator typologies: 3-clusters emerged for
males - generally violent; situational violent; and non-violent; and 4-clusters’
emerged for females - depressed attention seeker; hostile controller; undifferentiated
controller; and non-violent. Two separate MANCOVA were conducted, one for
males and one for females, using age and SES as covariates, to test whether the
typologies differed by risk factors (dependent variables). Overall, adolescent
perpetrator typologies differed by distal and proximal risk factors. Typologies
differed from one another within their gender as well as between gender groups (e.g.
males vs. females). Generally violent males had higher scores on parent-child
violence and witnessing inter-parental violence, compared to situational violent and
non-violent males. Generally violent males had higher scores on acceptance of
violence, school/community violence, alcohol/drug use, and number of dating
partners compared to non-violent males. Compared to non-violent females, the
depressed attention seeker females had higher scores on school/community violence,
seriousness of relationship, number of dating partners, and length of time dating. The
hostile controller females and undifferentiated controller females had higher scores
on parent-child violence, alcohol/drug use, and number of dating partners compared
to non-violent females. This study suggests a need for clinical findings to be

xi
included in initial assessments and treatment matching for adolescent perpetrator
typologies.

1
Chapter 1: Introduction
Adolescent dating violence may be the beginning of a progressive cycle of
violence that continues into adulthood if left unchecked (O’Leary, Malone, & Tyree,
1994). The elevated prevalence rates of dating violence highlight the impact of this
problem. Thirty-five percent (35%) of adolescents report some mild form of abuse,
such as threats, pushing, or slaps, and 10-20% of youth report more severe forms of
violence, such as being hit, slapped, or forced to have sex (Carver, Joyner, & Udry,
2003; Centers for Disease Control, 2000; & Coker et al., 2000). Overall rates of
dating violence among high school students have been found to be as high as 57%
(Cascardi, Avery-Leaf, O’Leary, & Smith-Slep, 1999). Dating violence occurring in
mid-adolescence (ages 13-15) may become a pattern that repeats and escalates into
adulthood (O’Leary, Malone, & Tyree, 1994). This is consistent with adult intimate
partner violence research, which found that the severity of violence increased the risk
for the victim if the pattern was established in adolescence (Feld & Strauss, 1989).
This is particularly problematic, because as the pattern evolves, the risk of
injury or even death increases with age. Thirteen thousand U.S. women 18 and older
lose their lives each year to intimate partner violence, and the costs of intimate
partner violence exceed $5.8 billion each year (National Center for Injury Prevention
and Control, 2003). Although in the past few decades researchers have begun to
investigate how and why adolescents use violence against their dating partners,
researchers have not yet examined possible typologies of adolescent perpetrators of
dating violence. The examination of specific characteristics or profiles of adolescent

2
perpetrators may be helpful in identifying and understanding subgroups or
differences among adolescent perpetrators, as well as helping tailor treatment to
these specific characteristics.
This study sought to identify adolescent dating violence perpetrator
typologies, and differences among them. In order to accomplish this task, the study
reviewed historical foundations of adult intimate partner typology research, as well
as existing adolescent dating violence research, including associated risk factors. A
theoretical basis for adolescent dating violence perpetrator typologies was also
explored. Specifically, the study discussed social learning theory and attachment
theory as they relate to the subject. Further, two cluster analyses were performed—
one for males, one for females—to identify adolescent perpetrator typologies.
Research objectives were developed from the literature review and these analyses.
Methodology included discussions of participants in the study, data collection
procedures, instruments, and preliminary findings. Results of this study included a
review of the findings of the statistical analyses (used to test hypotheses). Lastly, the
discussion elaborated on the results, putting them in the context of existing research
and theory in the field.

3
Chapter 2: Foundations of Typologies in Adolescent Perpetrators
of Dating Violence

Historically, researchers investigating violence in intimate partner
relationship research mainly targeted adult male batterers, who were thought to be a
homogeneous group in terms of the aims of such research and the problem of battery
being investigated. The researchers would average scores on measures of interest
across the entire sample of violent husbands, and then compare the mean score of the
violent sample with that of a non-violent sample. The problem with this procedure
was that it did not fully articulate the significant differences between the violent–
non-violent groups. Consequently, this approach resulted in researchers discounting
important variables that existed within the groups. For example, Neidig, Collins, &
Friedman (1986) were unable to distinguish batterers and non-batterers on measures
of attitudes toward women. Saunders (1992) argued that the variability of such
attitudes among violent men might be the reason no difference was found between
groups. Saunders (1992) found that scores on measures of attitudes toward women
were distributed bi-modally. For instance, one group was found to have liberal
attitudes, while the other was found to have more conservative attitudes. Similarly,
Holtzworth-Munroe & Stuart (1994) found in their review that violent husbands
varied along a number of important dimensions, such as severity of violence, anger,
depression, and alcohol abuse.
These findings supported the need for a reliable and valid typology to
increase understanding and identify different underlying processes leading to

4
intimate partner violence. Researchers reasoned that by comparing various subtypes
of violent groups, they might be able to clarify how each type differed from non-
violent groups, as well as from each other. They speculated that typology research
would allow researchers to systematically analyze how and why violent groups
differed in their use of violence against their partners. They felt that typology
research might not only increase understanding of such perpetrators, but also
increase treatment effectiveness through better patient–treatment matching.
Historically, most treatment programs for batterers were standardized and applied to
all violent men without consideration of their differences. However, researchers have
suggested that certain treatments might be more effective for one subtype over
another, and that matching subtype with treatment might lead to improved outcomes
in therapy (Gandolf, 1988; Saunders, 1992).
Historical perspective of typology research in intimate partner violence
Early typology research relied on two methods of analysis: rational/deductive
and empirical/inductive strategies. The rational/deductive method focused on
researchers’ clinical observations (e.g. Elbow, 1997). Within this approach, batterers
were split into groups on the basis of a priori theoretical assumptions and compared
on available data (e.g. Hershorn & Rosenbaum, 1991). Several typologies used this
method. For example, Faulk (1974) proposed five types of male batterers:
stable/affectionate, dependent/passive, dependent/suspicious, dominating, and
violent/bullying. Elbow (1977) identified four groups of batterers: the approval
seeker, the defender, the incorporator, and the controller. Caesar (1986) used the

5
Minnesota Multiphasic Personality Inventory (MMPI) with 26 batterers and found
three types: altruist, rescuer, and tyrant. Hershorn and Rosenbaum (1991) also used
an MMPI-derived scale measuring over-controlled hostility to separate 41 batterers
in treatment into two subgroups. One group was composed of violent husbands who
over-control their hostility (i.e., do not express anger until it summates and
explodes), and the second group was composed of those who under-control hostility
(i.e., have few controls against the expression of angry and aggressive impulses).
The researchers found the over-controlled hostile group engaged in more severe
marital violence compared to the under-controlled group, who engaged in less severe
but more frequent violence, used more aggression outside of the marital relationship,
and were generally more hostile.
Empirical/inductive methods, on the other hand, relied on factor analysis or
cluster analysis to identify subgroups of batterers. For instance, Gandolf (1988) and
Saunders (1992) used cluster analysis to examine what subtypes of male batterers
emerged from available data, and then compared the resulting clusters on other
external variables. Cluster analysis has been described as a statistical method that
“identifies and describes groups of individual cases defined by similarities along
multiple dimensions of interest. These groupings can form the basis for
understanding normal development, risk, or other outcomes” (Henry, D.B., Tolan,
P.H., and Gorman-Smith, D., 2005, p.1). Saunders (1992) performed a cluster
analysis on self-report data gathered from 165 married men entering treatment for
spousal abuse. Saunders selected six variables from the literature and entered them

6
into a cluster analysis: depression, anger, generalized violence, severity of marital
violence, attitudes toward women, and alcohol use during violent incidents. Three
aggressor subtypes emerged: family-only, emotionally volatile, and generally
violent.
Types of intimate partner violence typologies
By far the most comprehensive domestic violence typology research was
Holtzworth-Munroe and Stuart’s (1994) model based on clinical observations, a
priori hypotheses, and inductive approaches (e.g. factor or cluster analysis).
Holtzworth-Munroe and Stuart’s model consists of three theoretical levels. First, the
social-cultural level posits that marital violence exists because today’s violent and
patriarchal society encourages the use of physical aggression in families and the use
of violence to dominate women. Second, interpersonal theories sustain the middle
level, such as family systems models, that view the etiology of marital violence in
family interactions. Third, the individual or intrapersonal level uses many theories
that explain why some individuals engage in physical aggression, and others do not.
Holtzworth-Munroe and Stuart (1994) chose to integrate several different
intrapersonal theories into the developmental model of perpetration of marital
violence, such as family of origin violence, social skills, attachment, attitudes toward
women, and attitudes toward violence. Based on a review of the literature,
Holtzworth-Munroe and Stuart (1994) first proposed that three subtypes (family
only, borderline-dysphoric, and generally violent-antisocial) would be identified
using three descriptive dimensions (severity of marital violence, generality of

7
violence, psychopathology), and would differ on distal and proximal correlates of
violence. Later, Holtzworth-Munroe, Meehan, Herron, Rehman, and Stuart (2000)
tested this batterer typology by recruiting 102 male batterers and their wives from the
community along with two comparison groups of non-violent couples (i.e., maritally
distressed and non-distressed). Four clusters of violent men were identified: family-
only, borderline-dysphoric, generally violent-antisocial, and low-level antisocial.
Once these groups were formalized, the researchers used analysis of variance to
examine possible risk factors for the differing subtypes based on distal-historical and
proximal correlates, which in prior research had been found to be important
correlates of intimate partner violence. Distal-historical variables were composed of
factors occurring in childhood, including witnessing parental violence, experiencing
child abuse, methods of child discipline, and associations with deviant peers.
Proximal variables were related to situational contexts, such as attachment-
dependency styles, impulsivity, social skills, attitudes toward women, and attitudes
toward violence.
Typologies for adults
In general, the four batterer groups differed, as hypothesized, on these
measures of distal and proximal correlates of violence. The borderline-dysphoric
group tended to have the highest levels of preoccupied and fearful attachment. This
group also had the highest levels on related constructs, such as dependency on
spouse and jealousy. However, they did not always differ significantly from other
groups on these measures. The borderline-dysphoric group also scored high on

8
impulsivity. The generally violent antisocial group had the highest scores on the
dismissing scale from the Relationship Style Questionnaire (RSQ: Bartholomew &
Horowitz, 1991), which is a 30-item measure of attachment styles in romantic
relationships. However, the group’s difference was not statistically significant. The
generally violent antisocial group also had high scores on measures of impulsivity
and the lowest response on competency in both marital and non-marital hypothetical
situations (i.e. skills). The borderline-dysphoric and generally violent antisocial
groups also tended to have the most accepting attitudes toward violence and most
hostile attitudes toward women. Again, as predicted, among the violent subgroups,
the family-only batterers tended to have the lowest reported levels of the proximal
correlates of violence and closely resembled the non-violent group. The low-level
antisocial group tended to score between the family-only and generally violent
antisocial group on many of these measures.
Several other typology models of intimate partner violence emerged around
this time. For example, Johnson (1995) provided a two-category spousal abuse
typology, namely, common couple violence and patriarchal terrorism. Common
couple violence consists of both partners using occasional violence toward one
another in response to everyday stimuli. According to Johnson, this type of violence
is not related to gender and rarely escalates into life-threatening or serious violence.
Common couple violence is not related to patriarchy, but rather is part of a culture
that accepts violence in general. Patriarchal terrorism is a form of "terrorist control"
of women by their male partners and ex-partners. This type of violence is directly

9
related to patriarchy. Patriarchy is the historical tradition of men's legal and social
"right" to control women. Patriarchal terrorism tends to be more serious, occur over
a longer period of time, and be more likely to escalate into life-threatening violence
than common couple violence. Johnson and Leone (2005) modified these terms by
renaming common couples violence and patriarchal terrorism to situational couples
violence and intimate terrorism, respectively. Situational couples violence is not
rooted in a general pattern of control, but rather occurs from a specific conflict
situation that escalates to violence. This definition suggests that the relational
violence is not embedded in a general pattern of controlling behavior. Intimate
terrorism is less about patriarchy and more about the perpetrator’s attempt to
dominate a partner and control the relationship. The domination includes a wide
range of power and control tactics, including violence.
Similar to Johnson (1995), Johnson and Leone (1995) and Prince and Arias
(1994) also found two types of battering typologies: violence related to high levels of
frustration, and violence based on personal preferences and convictions. These
researchers examined the interactions between self-esteem, the desire for control
over events in one’s life, and the degree of perceived personal control over important
life pursuits. They used these variables to predict abuse status in a sample derived
from local clinics and newspaper advertisements. Interestingly, two different patterns
were associated with domestic violence in their analysis. Results revealed two
subgroups of men at high risk for engaging in domestic violence: (1) men low on
self-esteem, low on desirability of control, and low on perceived personal control;

10
and (2) men high on self-esteem, high on desirability of control, but low on
perceived personal control. These findings indicate that self-esteem may influence
partner violence through complex interactions with the perception of, and desire for,
control.
Another interesting typology that received attention in the popular literature
was developed by Gottman, Jacobson, Rushe, Shortt, Babcock, La Taillade, & Waltz
(1995), and Jacobson & Gottman (1998), who studied husbands’ heart rate reactivity
during a marital conflict discussion to classify batterers. Gottman et al. (1995)
defined reactivity as the difference in heart rate between baseline condition and
average heart rate over the first 5 minutes of a subsequent 15-minute marital conflict
discussion. Men whose heart rates decreased from baseline were called Type I
batterers, or cobras, and men whose heart rates increased from baseline were labeled
Type II batterers, or pit bulls (Gottman et al., 1995; Jacobson & Gottman, 1998).
More recently, researchers have found similar two-type typology models.
Chase, O’Leary, and Heyman (2001) found a two-category typology of adult
perpetrators: reactive and proactive aggressors. A selection system for categorizing
partner-violent men as either reactive or proactive aggressors was developed and
evaluated in the study. Reactive violence was characterized as a response (accurately
or inaccurately) to perceived threats or frustrations in the context of high affective-
physiological arousal and minimal cognitive processing. Proactive violence was
characterized as planned, methodical, and goal-oriented behavior, enacted in a
context of minimal-to-decreasing emotional and physiological arousal. Tweed and

11
Dutton’s (1998) empirical research found similar two-type typologies: Type I
(impulsive) and Type II (instrumental). Impulsive perpetrators were found to have
personality traits associated with increased reactivity to perceived partner rejection
or betrayal. Instrumental perpetrators had personality profiles indicating a readiness
to use aggression to get interpersonal needs met in a relationship (i.e. antisocial
tendencies, narcissism, and aggression) (Tweed & Dutton, 1998).
The typologies derived from these studies used either one or more of the
following variables either deductively or inductively to develop their models:
severity of violence, generality of violence, and psychopathology. Researchers used
these variables primarily for selecting or empirically deriving (factor or cluster
analysis) different subtypes or profiles of batterers.
In summary, past intimate partner violence research has focused largely on
adult male samples. Researchers generally believed these samples to be
homogeneous, until no group differences were found when compared to non-violent
groups, when examining the following key variables: attitudes toward women,
severity of violence, anger, depression, and alcohol abuse. These findings supported
the need for typology research, which has increased in the past two decades. Two
methods have been used to advance typology research: rational/deductive and
empirical/inductive. Several typology models have been developed (Johnson, 1995;
Prince and Arias, 1994; Chase, O’Leary, and Heyman, 2001; and Tweed and Dutton,
1998). Additionally, Holtzworth-Munroe and Stuart’s (1994) adult typology model
was empirically tested (Holtzworth-Munroe, Meehan, Herron, Rehman, and Stuart,

12
2000), and the three hypothesized subtypes of batterers, as well as a fourth subtype,
were identified and found to be statistically significant in validating their model.
Finally, adult typologies of male batterers were largely derived using three key
variables or dimensions: severity of violence, generalized violence, and
psychopathology.
Typologies for adolescents
There has been a noticeable lack of research on typologies of adolescent
perpetrators of dating violence, for many possible reasons. First, the fields of
typology research and adolescent dating violence research are still very young.
Researchers are just starting to understand the intricacies of adult intimate partner
violence. In addition, the complexity of adolescent perpetrator typology models may
have also contributed to a lack of progress in this area due to the rapid turnover in
teen relationships, peer influence and family pressures, and an increase of bilateral
conflict in the adolescent relationship (O’Leary & Slep, 2003). Furthermore, it may
be difficult to match existing adult typology models because adult development is
largely fixed, whereas adolescents are still growing and developing. Furthermore,
adolescent dating violence researchers may lack familiarity with adult batterer
research, and vice versa.
One early study (Stith, Jester, and Bird, 1992) did examine typologies of
dating violence and may provide a bridge between adult and adolescent models.
Using cluster analysis, these researchers derived subtypes of male and female college
students who had used violence in their dating relationships. This study differed from

13
adult studies in two important ways. First, Stith et al. (1992) examined dating
violence among college students rather than violence in the context of a marital-type
relationship. Second, in their sample, they included both men (n=69) and women (n
= 97) who had used physical aggression. In their cluster analysis, Stith et al. (1992)
entered measures of relationship functioning (e.g. love, relationship maintenance
activities, relationship conflict, and ambivalence about the relationship), relationship
negotiation styles, and general coping strategies. Four clusters were found: secure
lover, stable minimizer, hostile pursuer, and hostile disengaged. The four subtypes
were then compared to variables of theoretical interest, such as severity of violence,
level of emotional and verbal abuse, self-esteem, and length of relationship. Secure
lovers report the most love and the least amount of conflict of the four subtypes.
They tend to use a direct negotiation style with their partner and report less frequent
use of avoidant coping strategies. Secure lovers have higher self-esteem and longer
involvement with their partners. Physical and emotional abuse is rare and less severe
than the other clusters. Stable minimizers report moderate levels of love and
experience mid-range levels of conflict with their partners. The coping strategies
generally used by this cluster are avoidance and denial of problems. This group also
reports low levels of emotional and physical abuse. Hostile pursuers reported the
highest level of emotional abuse and moderate levels of physical abuse. This group
has a great amount of ambivalence toward their relationship, and these individuals
often use direct and indirect styles of negotiation. The last group, hostile disengaged
individuals, reports the most frequent and severe physical abuse and the lowest

14
feelings of love toward their partners. Their relationships are characterized by a
shorter duration and the highest amount of conflict among the four groups.
Interestingly, the researchers did not separate their findings by gender, and
the study was criticized because the perpetration of violence, and thus profiles of
perpetrators, may be fundamentally different and functionally dissimilar for men and
women. Also, the researchers did not compare the groups on measures of generality
of violence and psychopathology. Stith et al.’s (1992) unit of analysis was the dating
relationship, while adult typology researchers’ unit of analysis has been the batterer.
Risk factors associated with dating violence
Risk factors are essential to understanding the difference between perpetrator
subtypes (Holtzworth-Munroe, et al., 2000). Risk factors for dating violence may be
defined as “attributes of characteristics that are associated with an increased
probability of [its] reception and/or expression” (Hotalling & Sugarman, 1990, p.1).
Risk factors are correlates of dating violence and not necessarily causative factors.
Consequently, these variables may help to differentiate individuals involved in
dating violence from those who are not, but the variables can also be used as
symptoms or outcomes with implications for treatment. Just as risk factors gave rise
to adult batterer typologies in the adult batterer literature, the past study of
adolescent risk factors will also facilitate the work on subtypes of adolescent
perpetrators.
Riggs & O’Leary’s (1989) model provides a way of organizing the numerous
risk factors. These researchers developed an extensive model that provides an

15
overview of findings related to variables expected to be important for defining the
adolescent perpetrators’ subtype groups (See Figure 1). Riggs and O’Leary use two
theoretical levels to explain their risk factor model: contextual and situational.
Contextual variables represent more distal variables, and are composed of four
variables of theoretical interest to the adolescent perpetrator subtype groups:
witnessing inter-parental violence, parent-child physical aggression, acceptance of
violence, and exposure to school/community violence. Situational, or proximal
variables, are more predicative of when dating violence might occur, such as
alcohol/drug use, seriousness of relationship, number of dating partners, and length
of time dating.

Figure 1. Riggs and O’Leary’s (1989) Dating Violence Model

16
Witnessing inter-parental violence and parent-child physical aggression may
be important variables in differentiating adolescent perpetrator subtypes. Researchers
found adolescent perpetrators of dating violence were more likely to have
experienced child abuse or neglect (McCloskey, Figueredo, & Koss, 1995; Wekerle
& Wolfe, 1998; Wolfe, Werkele, Reitzel-Jaffe & Lefebver, 1998) and were more
likely to have witnessed inter-parental violence. Exposure to violence in the home is
an important variable within adolescent perpetrator typology research. With regard to
subtypes, Dodge, Lochman, Harnish, Bates, & Pettit (1997) found that reactive
aggressor subtypes of intimate partner violence experienced more physical abuse
from parents and witnessed less inter-parental violence than did proactive aggressors.
Direct violence from parents is likely to cause a disruption in the secure attachment
relationship with the parent. Therefore, the borderline/dysphoric group is likely to
have higher scores compared to other groups using parent-child aggression as a
dependent variable. The generally violent group is likely to have higher scores on
interpersonal violence because this group is known to use more severe violence. It is
likely that they learned this behavior from parental modeling in the home.
Acceptability of using violence in relationships has been found to be
associated with inflicting violence in a dating relationship (Malik et al., 1997;
O’Keefe, 1997; Tontodonato & Crew, 1992). The Dysphoric group and the generally
violent group are also likely to have higher scores on the acceptability of using
violence variable. Holtzworth-Munroe et al. (2000) used acceptability of using
violence to differentiate between subtype groups. They found the

17
borderline/dysphoric group and the generally violent group were more accepting of
violence compared to the family-only group. Exposure to school/community
violence has also been associated with the perpetration of dating violence in the non-
typological research (Malik et al., 1997). There are presently no adult typology
studies that have examined how the risk of school/community violence may
differentiate typology groups. However, it is likely that the generally violent group,
which engages in violence outside the family system, may be exposed to more
violence in the community by their association with violent peer groups.
Riggs and O’Leary (1989) define situational variables that are more
immediate and proximal correlates, and include variables that are predictive of when
and in what situations the aggression might occur. With regard to the present study,
four situational variables are thought to differentiate between subtypes:
alcohol/drugs, seriousness of relationship, number of dating partners, and length of
time dating. The generally violent group in this study is likely to have higher scores
on alcohol/drugs as a dependent measure. Alcohol and drug use is also an important
variable for differentiating between batterer subtype groups. Holtzworth-Munroe et
al. (2000) found that generally violent batterers are the most involved in substance
abuse compared to all other groups in the study.
The seriousness of the relationship, length of time dating, and the number of
dating partners may be particularly important variables for examining different
subtypes within the adolescent perpetrator typology model. Descriptive studies
indicate that dating violence tends to occur during the serious dating phase of a

18
relationship (Cate et al., 1982; Henton, Cate, Koval, & Christopher, 1983; Sigelman
et al., 1984). Consistent with prior studies on dating violence in college samples, the
length of time dating and number of dating partners was positively associated with
inflicting dating violence (Stets & Pirog-Good, 1987). The borderline/dysphoric
group is likely to have higher scores on the seriousness of the relationship variable
compared to other groups. This group has a tendency to be overly dependent and
preoccupied with the relationship, and likely to view it as serious. The
borderline/dysphoric group may have higher scores compared to other groups on the
measures of length of time dating and lower scores on the number of dating
relationships. The more dependent the dysphoric perpetrator is on the relationship,
the more likely violence will arise in that relationship.
Risk factors for females
The literature on risk factors for female perpetrators of intimate partner
violence is scant compared to the extensive literature on male perpetrators. As
indicated earlier, Holtzworth-Munroe and Stuart (1994) found three types of male
batterers: family-only, borderline/dysphoric, and generally violent/antisocial. Based
on their findings, Babcock et al. (2003) studied 52 women referred to a treatment
agency for abusive behavior. The women in their sample were categorized into two a
priori categories: partner-only women, who only reported using aggression against
their partners since age 18, and generally violent women, who reported using
violence in various circumstances since age 18. The researchers found that the
generally violent group reported more fights with different people and committed

19
significantly more physical and psychological abuse, and also inflicted more injuries
against their partners during the past year, compared to the partner-only women’s
group. There were no significant differences between partner-only and generally
violent groups’ experiences of abuse by their partners. Consistent with this author’s
hypotheses, generally violent women reported more current trauma symptoms, a
desire to hurt themselves, a desire to hurt others, and interpersonal problems
compared to the partner-only women’s group. Both the generally violent and partner-
only groups reported high rates of childhood physical and sexual abuse. Lastly, the
generally violent group reported seeing their mothers use violence against their
fathers more than the partner-only women’s group.
Henning, Jones, and Holdford (2003) studied a large adult sample of male
(2,254) and female (281) domestic violence offenders to examine demographic,
childhood family functioning, and mental health characteristics. This study did not
identify perpetrator subtypes or attempt to categorize subjects into groups based on
partner-only violence versus generally violent. Rather, the study measured the
different risk factors between males and females. Researchers found few differences
between the demographic characteristics of men and women arrested for domestic
violence. Women differed in that they were more likely to have attended college and
less likely to have worked outside the home. According to the WAIS-R, both men
and women had low IQs ranging from borderline intellectual functioning to mental
retardation. A similar trend was observed when comparing men’s and women’s
childhood experiences in relation to adulthood adjustment difficulties or

20
psychopathology (i.e., physical abuse, inter-parental physical aggression, parental
criminal behavior, or substance abuse). Men reported more corporal punishment by
primary caregivers, while women reported witnessing more severe inter-parental
abuse. Gender differences were also evident in the subjects’ mental health histories
and current mental health status. Male offenders reported more prior treatment for
substance abuse/dependence and were rated high risk for substance dependence and
childhood conduct problems prior to age 16. They also had a desire to continue the
relationship with the victim. Women offenders, on the other hand, reported being
prescribed more psychotropic medication, with a prior history of a suicide attempt.
Both men and women were equally likely to report significant distress in the sample.
Lastly, Henning et al. (2003) found symptoms of personality dysfunction and mood
disorders more prevalent in women arrested for domestic violence compared to their
male counterparts in the study.
Another study that focused on gender differences was that of Luthra and
Gidycz (2006), who empirically tested Rigg’s and O’Leary (1989) model of dating
violence with a split sample of 100 female and 100 male college students. Students
completed assessments concerning the occurrence of violence in their dating
relationships. Multivariate logistic regression analyses were performed to determine
the most salient predictors of dating violence for each gender. Findings revealed that
the model was more accurate in predicting female perpetration of dating violence, as
compared with male. The model accurately classified 83% of violent women and
only 30% of violent men. A partner’s use of aggression was the single largest

21
predictor of female dating violence perpetration in the study. Women were 108 times
more likely to perpetuate violence if they were the recipient of physical violence
from a dating partner, while men with violent partners were only 1½ times more
likely to perpetrate violence. Researchers suggest that women inflict violence as a
response to dating violence.
The next largest predictor of female perpetration of dating violence was
alcohol. Women and men who reported alcohol use were equally likely (five times)
to perpetrate violence against their dating partner. Therefore, alcohol use is a risk
factor for dating violence for both men and women.
The study provided mixed support for the idea that violence was transmitted
across generations. Women who had violent fathers were almost three times more
likely to perpetrate violence against their dating partners. However, parental violence
was not related to men perpetrating dating violence. This suggested that witnessing
parental violence might have a greater impact for women than for men. Relationship
length was only a significant predictor for male perpetration of violence, not female.
Men were twice as likely to perpetrate violence with each 6-month increase in the
duration of the relationship. Several studies have consistently shown that dating
violence is linearly related to relationship length (Lewis & Fremouw, 2001; Ray &
Gold, 1996), and that the severity of abuse increases over time (Billingham, 1987).
There are few teen studies examining female adolescent perpetrators of
dating violence to date. Banyard, Cross, and Modecki (2006) surveyed a subset of
980 adolescents ages 11 to 19, as part of the Cooperative Extension to survey all

22
youth in target communities about risky health behaviors. The study examined
correlates of perpetration of either physical or sexual abuse across all levels of the
ecological model (individual, family, and community factors). The researchers used
exploratory analyses to examine patterns of significant risk correlates separately for
males and females, using logistic regression with a composite perpetration variable
as the outcome. The study found that for females, the only significant variable was
depressed mood, with a trend for increased odds of perpetration if parents were
divorced. Male perpetrators were found to be at risk for alcohol and drugs, as well as
being a victim of sexual abuse with a trend toward depressed moods. The variables
revealed a greater variance in perpetration for boys than for girls. Further research is
needed to examine typology groups, adolescent girls who are perpetrators of dating
violence (using separate samples), and related risk factors for dating violence among
these groups.
In summary, there has been a noticeable lack of typology research on
adolescent perpetrators of dating violence, especially among adolescent girls.
Typology and adolescent dating violence research are still in their infancy.
Adolescent perpetrator typology is complex due to developmental issues and
difficulty combining adult and adolescent research. Stith, Jester, and Bird (1992)
examined dating violence in a college sample of men and women, and found four
clusters: secure lover, stable minimizer, hostile pursuer, and hostile disengaged. The
study did not separate by gender, nor did researchers compare groups on measures of
generality of violence or psychopathology. Risk factors or correlates of dating

23
violence were discussed. Riggs & O’Leary’s (1989) model was discussed as a way
of organizing the risk factors into two groups: distal (witnessing inter-parental
violence, parent-child physical aggression, acceptance of violence, and exposure to
school/community violence) and proximal (alcohol/drugs, seriousness of
relationship, number of dating partners, and length of time dating). Risk factors for
females were also reviewed. Babcock et al. (2003) categorized women offenders into
a partner-only violent group and generally violent group. The generally violent group
reported more fights with different people, more psychological abuse, and more
injuries inflicted on their partners compared to the partner-only group. Further, the
generally violent group had higher rates of trauma symptoms, desire to hurt
themselves or others, witnessing their mother using violence against their father, and
more interpersonal problems, compared to the partner-only group.
Henning et al. (2003) also did not research typologies, but focused on
different risk factors among males and females. Researchers found few differences
between men and women in a large domestic violence offender study. Women who
attended more college and worked less outside the family were linked to more
intimate partner violence. Both men and women offenders had low IQ’s and reported
significant distress. Men reported more corporal punishment from primary
caregivers, prior treatment for substance abuse/dependence, higher risk for substance
dependence, and greater desire to continue a relationship with the victim. Women
offenders reported more witnessing of severe interpersonal abuse from primary

24
caretakers, more prescription of psychotropic medication with a history of a suicide
attempt, personality dysfunction, and mood disorders.
Luthra and Gidycz (2006) also studied a college sample, finding women were
108 times more likely to perpetuate violence if they were the recipients of physical
violence compared to 1½ times for men. Women and men were equally likely to use
alcohol when perpetrating dating violence. Women with violent fathers were 3 times
more likely to perpetuate violence against their dating partners. Relationship length
was only significant for male perpetrators of dating violence. Banyard, Cross, and
Modecki (2006) conducted one of the few studies to examine adolescents ages 11 to
19. Depressed mood was the only significant variable found for girls perpetrating
dating violence, with a trend toward divorced parents. Males were found to be at risk
for alcohol/drug use and sexual abuse, with a trend toward depressed mood. More
research is needed to examine typologies among female as well as male adolescent
perpetrators of dating violence along with related risk factors in order to further
understand the complexity of dating violence. Theoretical models of adolescent
perpetrator typologies of dating violence may also be helpful to further our
understanding of dating violence perpetrator behavior.

25
Chapter 3: Theoretical Foundations and Research Objectives
Theoretical frameworks supporting this study
Social learning theory provides a theoretical foundation that may enable us to
understand why batterer groups differ from one another. Social learning theory
proposes that individuals acquire novel behaviors and expand their behavioral
repertoires by observing others’ behavior. This model suggests that observational
learning is more important in shaping new behavior as is classical and operant
conditioning (Bandura, 1969, 1973). Individuals watch others and then use imitation
to perform novel behaviors. Bandura also asserts that individuals can learn new
behaviors without performance and without reinforcement. Simply stated, an
observer may copy a model’s behavior long after he or she saw the action performed.
The most commonly cited mechanism for the intergenerational transmission
of spouse abuse is social learning. That is, observing physical aggression in parents
provides a model by which children acquire violent behaviors (Bandura, 1969, Corvo
& Carpenter, 2000, MacEwen, 1994, Widom, 2000). Social learning theory provides
a critical explanation for why some adolescent perpetrators of dating violence do not
learn the skills necessary to resolve relational conflict without resorting to aggression
in their dating relationship. Essentially, adolescents do not learn conflict resolution
because they do not have models to provide them with novel or observed behaviors
where conflict is negotiated or resolved in a non-violent way. Bandura (1973) argues
that growing up in an abusive family teaches individuals that the use of aggression is

26
a viable means for dealing with interpersonal conflicts and increases the likelihood
of becoming involved in future aggression.
The disruption of the development of social skills by aggression in the home
may provide some explanation for why batterer groups differ from one another by
subtype. In addition, several classic social learning theory studies may provide
support for why adolescent dating violence perpetrator subtypes resemble adult
intimate partner violence typology subtypes, such as common couples,
borderline/dysphoric, and generally violent. The studies found that male models
were more likely to be imitated than were female models, and the effects of males’
modeling were more enduring over time (Bandura, Ross, & Ross, 1961; Bussey &
Bandura, 1984; see also Hicks, 1965). This suggests that a child growing up in a
home where the father models family-only violence or generally violent aggression
may be imitated by the child in the future.
Social learning theory may not be sufficient to explain evidence of some
processes occurring beyond the imitation of specific models of violent behavior
(Corvo & Carpenter, 2000). Some researchers have theorized that the process
underlying the intergenerational transmission of behaviors such as family violence is
not modeling, but rather attachment style. Attachment theory provides rich
implications for understanding typologies of dating violence perpetrators, and has
been used to explain adult batterers’ violent behavior. Bowlby (1988) suggested that
a violent husband’s insecure attachment to his partner leads to fear of losing her,
which in turn may lead to violent behaviors in the relationship. Several other

27
researchers have examined the link between attachment and intimate partner
violence. The researchers found that violent batterers tended to report a disrupted
attachment to their partners, an over-dependence on their partners, or a
preoccupation with their partners (Dutton, Saunders, Starzomski, & Bartholomew,
1994; Murphy, Meyer, & O’Leary, 1994).
Male batterers were also found to display more attachment problems than
non-violent men, and a variety of insecure attachment patterns were found to be
associated with a greater likelihood of violent behaviors (Babcock, Jacobson,
Gottman, & Yerington, 2000; Bowlby, 1988; Dutton et al., 1994; Holtzworth-
Munroe, Stuart, & Hutchinson, 1997; Murphy et al., 1994). Dutton et al. (1994)
found that composite attachment scores (anxious and avoidant) were significantly
correlated with other abuse-related characteristics, including borderline personality
organization, anger, jealousy, and trauma symptoms. One explanation for
perpetrators of dating violence may involve the negative psychological consequences
of insecure-avoidant attachment (Ainsworth, 1989). Exposure to violence in the
home can undermine the development of children’s abilities to trust others, which in
turn makes it difficult for them to form a secure attachment relationship (Janoff-
Bulman, 1992).
Attachment style is related to the perpetration of dating violence because
individuals that are insecurely attached (preoccupied/anxious-ambivalent or
avoidant) are more likely to respond with violence due to frustration in the
relationship. In a preoccupied (anxious-ambivalent) individual, the violence could be

28
a reaction to criticism and/or threats to terminate the relationship. The violence
would be considered a futile attempt to re-establish the relationship (Pistole &
Tarrant, 1993). Feeney and Noller (1992) assessed attachment style in an
undergraduate population upon relationship dissolution. The anxious-ambivalent
group reported being the most upset and more surprised at the dissolution of the
relationship. This group is highly dependent on their partner in order to maintain
positive self-regard (Feeney & Noller, 1990; Hazan & Shaver, 1987). Therefore, it is
likely that this group will have a difficult time adjusting to the anxiety and
ambivalence in the relationship due to insecure attachment.
The study suggests that individuals with anxious-ambivalent attachment
styles are more likely to experience relationship instability and dissolution (Feeney
& Noller, 1992). Research suggests that insecure attachment in men is strongly
related to violence within intimate relationships (Dutton et al., 1994; Dutton &
Golant, 1995; Mayseless, 1991). Holtzworth-Munroe, Meehan, Herron, Rehman, &
Stuart (2000) define the borderline/dysphoric group as having difficulty forming a
stable and trusting attachment with an intimate partner. They are highly dependent
on, but fearful of losing, their partners. They are jealous, lack adequate marital
relationship skills, and have hostile attitudes toward women that moderately support
violence. This often leads to depressive thoughts and feelings.
Integrating adult typologies with adolescent typologies
The present study sought to examine profiles for adolescent male and female
perpetrators of dating violence using a separate analysis. The same variables used to

29
construct profiles for adolescent males may also be used for adolescent female
perpetrators. Existing adult typology research provided some insight into identifying
cluster dimensions and subtypes for present cluster analysis of adolescent
perpetrators of dating violence. The non-violent subtype group usually acts as a
comparison group. This is consistent with Holtzworth-Munroe et al. (2000), who
also used a non-violent comparison group. The family-only violent partner group
was compared to the other subtype groups as well as non-violent comparison group;
these groups are usually very similar. Stith et al.’s (1992) stable minimizer group
appears to be similar to the family-only group in that they are both likely to have
lower scores of emotional and physical abuse compared to other groups in each
study. The situational violent group in adolescent dating focuses less on control and
more on conflict situations that lead to dating violence. Therefore, it seems to be a
more appropriate term for dating violence among adolescents. Although a number of
the dimensions that Stith et al. used were not used in the present study, it is expected
that the situational violent group profile in this study will also have lower scores
compared to other groups on severity of dating violence, generality of violence, and
internalizing and externalizing psychopathology. Stith et al. also found some
relationship satisfaction and conflict with the stable minimizer group compared to
other subtypes. The situational violent group may have similar results on these
cluster variables.
No typology studies to date have looked at context of violence variables.
Hickman, Jaycox, & Aronoff (2004) report a need for empirical support on reaction

30
and reason for dating violence. Adolescent perpetrators of dating violence may react
to violence very differently. They may also use violence against a partner for
different reasons. Moreover, it is likely that the situational violent group will react to
violence with anger and report attention seeking as the primary reason for using
violence.
A dysphoric group is also likely to evolve from the current cluster analysis of
this adolescent sample, and it is expected to closely resemble the Holtzworth-
Munroe et al. (2000) findings for their borderline/dysphoric group, based on similar
cluster dimensions. Therefore, it is likely that the dysphoric group in the present
study would have higher scores on measures of severity of dating violence,
generality of violence, and internalizing psychopathology, with lower scores on
externalizing psychopathology. The dysphoric group is also likely to resemble Stith
et al.’s (1992) hostile pursuer group on similar cluster dimensions. Stith et al.
reported higher scores for this group on emotional abuse and physical abuse
compared to other subtypes. They reported a great amount of ambivalence toward
their relationship with direct and indirect styles of negotiation on similar cluster
dimensions. The dysphoric group in this study may report anger as well as fear and
sadness as the primary reaction to violence. The dysphoric group may also report
using violence for attention seeking reasons, such as difficulty expressing feelings,
attempts to prove love or get attention, or jealousy. The dysphoric group, in this
study, may have less relationship satisfaction and more conflict due to possible
ambivalence in the relationship.

31
The generally violent group, in this study, is expected to resemble the
Holtzworth-Munroe et al. (2000) findings for their generally violent subtype group,
compared to other subtype groups. Consistent with their findings, the generally
violent subtype group will have higher scores on severity of dating violence,
generality of violence, and externalizing psychopathology, and lower scores on
internalizing psychopathology. Generally violent perpetrators may report little or no
emotional reaction to using violence. Or, they may report controlling reasons for
violent behavior, such as to show anger, feel more powerful, get control over their
partner, get back at their partner, displace anger (felt toward someone else), or to
punish their partner. They may also rationalize their violent behavior as a result of
drug use. This generally violent group is also likely to have lower satisfaction and
higher relationship conflict. These expectations may also resemble the hostile
disengaged subtype groups found in Stith et al.’s (1992) study with similar subtypes
and cluster dimensions. The hostile disengaged subtype group reported the most
frequent and severe physical abuse and lower feelings of love toward their partners.
Their relationships are characterized by shorter durations and the highest amount of
conflict among the four groups. Therefore, this study’s generally violent subtype
group is likely to have higher scores on frequency and severity of violence measures,
as well as the conflict in the relationship measure, and the lowest scores on
relationship satisfaction.
Although the above anticipated findings are similar to those found in adult
male perpetrators, differences are also expected given this study’s much younger

32
population. The main expected difference between findings in the adult batterer’s
literature and the present study concerns the percentages expected in each subtype.
Holtzworth-Munroe et al. (2000) found 36% of the subjects fell in family only, 33%
in the low level antisocial, 15% in the borderline/ dysphoric, and 16% in the
generally violent antisocial group. This study is likely to have a much higher
percentage representing the situational violent group, maybe as high as 50 to 60%.
The dysphoric group is expected to have the next highest percentage. Lastly, it is
expected that a much lower percentage of adolescents will fall into the generally
violent group, possibly as low as 10%. This is expected given that prevalence studies
have shown that the severity of violence increases with age (Feld & Strauss, 1989).
Expectations and Barriers to Current Research
The proximal adolescent risk factor variables are expected to have findings
similar to those of the Holtzworth-Munroe et al. (2000) analysis of variance with
adult risk factors for intimate partner violence. Several variables in this study were
not included in the Holtzworth-Munroe et al. (2000) study, so it is difficult to predict
how proximal dependent variables in this study will differ between adolescent
perpetrator subtype groups. The generally violent group is likely to score higher on
alcohol/drug use compared to the other adolescent subtype groups. This is consistent
with the Holtzworth-Munroe et al. (2000) findings. The seriousness of relationship,
number of dating partners, and length of time dating variables were not examined in
the Holtzworth-Munroe et al. (2000) study. Therefore, predictions on these risk
factors are not clear. As indicated earlier, the dysphoric group is likely to have higher

33
scores on the seriousness of relationship variable compared to other groups, because
they are overly dependent and preoccupied with the relationship. They are also likely
to have higher scores on the measures of length of time dating, with lower scores on
the number of dating relationships, compared to other groups. The more dependent a
dysphoric perpetrator is on the relationship, the more likely violence will arise.
Research Objectives
Dating violence is a major social problem. It begins in mid-adolescence,
repeats, and escalates in severity into adulthood. Risk of injury, death, and costs
related to dating violence increase with age. A cluster analysis was done to identify
subtypes of adolescent male and female perpetrators of dating violence. A
MANCOVA design tested the null hypothesis of no significant group differences
across distal and proximal risk factors for adolescent dating violence. The
independent variable was adolescent perpetrator typology. The covariates were two
demographic variables (age, SES). The dependent variables were eight dating
violence risk factors. This study seeks to understand underlying factors that
contribute to this problem, focusing on the following hypotheses:
1. Distal risk factors: parent-child violence, witnessing inter-parental
violence, acceptance of violence, and school/community violence.
2. Proximal risk factors: alcohol/drug use, seriousness of relationship,
number of dating partners, and length of time dating.

34
Ho1: After controlling for demographic variables, there were no significant
differences between male and female adolescent perpetrator typologies on
distal and proximal risk factors.
Ha1: After controlling for demographic variables, there were significant differences
between male and female adolescent perpetrator typologies on distal and
proximal) risk factors.
Hypothesis for male sample:
Ho2: After controlling for demographic variables, the generally violent and
situational violent typologies will not score significantly higher on distal risk
factors than the non-violent typology group.
Ha2: After controlling for demographic variables, the generally violent and
situational violent typologies will score significantly higher on distal risk
factors than the non-violent typology group.
Ho3: After controlling for demographic variables, the generally violent typologies
will not score significantly higher on distal risk factors than the situational
violent male perpetrators group.
Ha3: After controlling for demographic variables, the generally violent typologies
will score significantly higher on distal risk factors than the situational
violent male perpetrators group.
Ho4: After controlling for demographic variables, the generally violent and
situational violent typologies will not score significantly higher on proximal
risk factors than the non-violent typology group.

35
Ha4: After controlling for demographic variables, the generally violent and
situational violent typologies will score significantly higher on proximal risk
factors than the non-violent typology group.
Ho5: After controlling for demographic variables, the generally violent typologies
will not score significantly higher on proximal risk factors than the situational
violent male perpetrators group.
Ha5: After controlling for demographic variables, the generally violent typologies
will score significantly higher on proximal risk factors than the situational
violent male perpetrators group.
Hypothesis for female sample:
Ho6: After controlling for demographic variables, the depressed attention seeker, the
hostile controller, and the undifferentiated controller violent typologies will
not score significantly higher on distal risk factors than the non-violent
typology group.
Ha6: After controlling for demographic variables, the depressed attention seeker, the
hostile controller, and the undifferentiated controller violent typologies will
score significantly higher on distal risk factors than the non-violent typology
group.
Ho7: After controlling for demographic variables, the hostile controller, and the
undifferentiated controller violent typologies will not score significantly
higher on distal risk factors than the depressed attention seeker violent female
perpetrators group.

36
Ha7: After controlling for demographic variables, the hostile controller, and the
undifferentiated controller violent typologies will score significantly higher
on distal risk factors than the depressed attention seeker violent female
perpetrators group.
Ho8: After controlling for demographic variables, the hostile controller typologies
will score significantly higher on distal risk factors than the undifferentiated
controller violent female perpetrators group.
Ha8: After controlling for demographic variables, the hostile controller typologies
will score significantly higher on distal risk factors than the undifferentiated
controller violent female perpetrators group.
Ho9: After controlling for demographic variables, the depressed attention seeker, the
hostile controller, and the undifferentiated controller violent typologies will
not score significantly higher on proximal risk factors than the non-violent
typology group.
Ha9: After controlling for demographic variables, the depressed attention seeker, the
hostile controller, and the undifferentiated controller violent typologies will
score significantly higher on proximal risk factors than the non-violent
typology group.
Ho10: After controlling for demographic variables, the hostile controller, and the
undifferentiated controller violent typologies will not score significantly
higher on proximal risk factors than the depressed attention seeker violent
female perpetrators group.

37
Ha10: After controlling for demographic variables, the hostile controller, and the
undifferentiated controller violent typologies will score significantly higher
on proximal risk factors than the depressed attention seeker violent female
perpetrators group.
Ho11: After controlling for demographic variables, the hostile controller typologies
will not score significantly higher on proximal risk factors than the
undifferentiated controller violent female perpetrators group.
Ha11: After controlling for demographic variables, the hostile controller typologies
will score significantly higher on proximal risk factors than the
undifferentiated controller violent female perpetrators group.

38
Chapter 4: Methodology
This chapter outlines the methodology used in this study to address the
research questions and hypothesis proposed in the last chapter. Presented first is the
proposed design, followed by a detailed discussion of the participants, data collection
procedures, and instruments analysis. The chapter concludes with a review of the
data analysis techniques used to address the study’s hypotheses.
Study Design
The principal goals of the study were to (a) identify subtypes of adolescent
male and female perpetrators of dating violence, and (b) examine possible risk
factors for differing adolescent male and female perpetrators of dating violence
subtypes based on distal and proximal correlates. To achieve these goals, the study
used a cross-sectional survey method, taking a cross-section of dating violence at one
point in time (Rubin and Babbie, 1993; Hulley & Cummings, 1988). Hulley &
Cummings (1988) suggested that cross-sectional survey studies are well suited to
examining associations and the plausibility of causal relationships between variables,
such as witnessing inter-parental violence and perpetrating dating violence. Thus, the
current study proposed a hypothesized model (Figures 2 and 3) for male and female
adolescent perpetrators of dating violence. The model identifies typology subtypes
and examines the relationship between distal and proximal risk factors and typology
subtypes based on Riggs and O’Leary’s (1989) dating violence model (Figure 1).

39

Figure 2. Hypothesized Model: Typologies of male adolescent perpetrators of
dating violence and differences among them. 1. Identify clusters; 2. Explore distal
variables’ relationship to clusters; and 3. Explore proximal variables’ relationship to
clusters.

Figure 3. Hypothesized Model: Typologies of female adolescent perpetrators of
dating violence and differences among them. 1. Identify clusters; 2. Explore distal
variables’ relationship to clusters; and 3. Explore proximal variables’ relationship to
clusters.

40
A MANCOVA design tested the null hypothesis of no significant group
differences across distal and proximal risk factors for adolescent dating violence.
The independent variable was adolescent perpetrator typology. Adolescent
perpetrator subtypes for males included generally violent, situational violent, and
non-violent comparison group. Adolescent perpetrator subtypes for females included
depressed attention seeker, hostile controller, undifferentiated controller and the non-
violent comparison group. The covariates were two demographic variables (age,
SES). The dependent variables were eight dating violence risk factors: witnessing
inter-parental violence, experiencing parent-child physical aggression, acceptance of
violence, school/community violence, alcohol/drug use, seriousness of relationship,
number of dating relationships, and the length of time dating.
Participants
Participants in the study included 972 high school students, 402 males and
570 females, ages 13-20. All participants met the following eligibility criteria: (a)
adolescent full-time students attending high school in the Los Angeles area; (b)
willingness to voluntarily participate in the study; (c) age less than 20 and at least 13
at the time of the study; (d) participants had a parent or legal guardian; (e) ability to
speak English; and (f) ability to read at the fifth grade level.
Data Collection Procedures
Data collection took place in 1997, part of a large dating violence study
(O’Keefe, 1997). Eight schools were asked to participate in the study. Two declined,
reporting that they did not permit researchers in their schools, or that the subject

41
matter might be too sensitive. The two schools that declined were composed largely
of families of middle- and upper-middle-class socioeconomic status. The six
participating high schools were in the Los Angeles area, including two inner-city and
three urban areas surrounding Los Angeles, and one in a more suburban area. Several
schools were located in inner-city high crime areas, whereas others were located in
lower crime areas.
Approval to conduct this research was obtained from the Institutional Review
Board of the University of Southern California. Data for the present analyses were
collected as part of a larger research study examining adolescent dating violence and
the effects of adolescents’ exposure to violence in different social settings (see
O’Keefe, 1997). The investigator distributed survey packets during regularly
scheduled class periods, with consent forms for both participants and their parents.
There was no threat of coercion connected to completing the survey. The
questionnaire was designed for comprehension at a fifth-grade reading level and was
pre-tested on a socio-economically diverse sample of adolescents. A total of 20-25%
in each class declined to participate or neglected to return parental consent forms.
The investigator was present during survey completion to answer students’ questions
and prevent discussion until all participants finished.
Completion of the questionnaire took about 45 minutes. After the survey was
completed, a researcher debriefed students on the purpose of the research and offered
information and available resources on dating violence. Fifty-five questionnaires
were eliminated because they were incomplete or incorrectly completed; another 33

42
questionnaires were excluded from analysis because participants reported they had
not started dating.
Instruments
A cluster analysis was done based on the literature review to explore whether
the subtypes of adolescent perpetrators of dating violence in this study resembled
existing typology research findings. The present study examined whether profiles
existed for adolescent male perpetrators of dating violence. A separate analysis was
conducted for females. However, since there were few prior studies examining
profiles of female perpetrators (even in the adult literature), initially the same
variables used for males were used to explore profiles of female adolescent
perpetrators. However, the initial cluster analysis, which used the same dimensions
for males and females, could not differentiate between female adolescent perpetrator
cluster groups. Therefore, the recipient of violence variable, a well known dating
violence risk factor for women, was added to the variables used in the initial cluster
analysis for the female sample to facilitate the clustering of typologies for female
perpetrators.
No a priori hypotheses were made for cluster analysis due to the exploratory
nature of the study. The following research questions were used to guide the cluster
analysis: Are there typologies in adolescent dating violence for males and females?
Do the typologies differ for males and females? In addition, a MANOVA for each
gender in which the variables used for clustering served as (temporary) dependent

43
variables was used to answer the question: What is the pattern of scores on the
clustered variables for each typology identified?
The descriptive dimensions were derived from the Holtzworth-Munroe et al.
(2000) study and were used in the cluster analysis to identify perpetrator subtypes.
The cluster analysis of subtypes of adolescent perpetrators of dating violence has
been exploratory at this point. Several groups may emerge in the cluster analysis:
situational violent, generally violent, and non-violent comparison group. Six cluster
dimensions were used to identify the adolescent perpetrator groups: severity of
dating violence, generality of violence, psychopathology, contextual factors of dating
violence (i.e. the reaction to physical aggression scale and the reason for physical
aggression scale), relationship satisfaction, and relationship conflict.
Cluster Dimensions
Severity of Dating Violence. Severity of violence will be measured in this
study using a modified version of the violence subscale of the Conflict Tactics Scale
(CTS) (Straus, 1979) was used to assess the amount of dating violence ever inflicted
on a dating partner. Respondents indicated how often they had ever used the
following acts of physical violence during the course of their dating: threw an
object; pushed, grabbed, or shoved; slapped, kicked, bit, or punched; hit or tried to
hit with an object; beat up; threatened with a knife or gun; used a knife or gun; or
forced sexual activity. Responses were made on a 7-point scale ranging from 0 =
none to 6 = more than 10 times. Scores were summed to provide a total inflicted

44
dating violence score. The Cronbach alphas for this sample were separated for both
males (.77) and females (.79).
Psychopathology. Psychopathology will be measured using the internalizing
and externalizing measures from the Youth Self-Report (YSR). The YSR is a 113-
item instrument that has been widely used to assess adolescent behavioral problems.
This measure yields total broad-band problem scores and narrow-band subscale
scores assessing internalizing and externalizing mental health symptoms. This study
will focus on the internalizing (e.g. depression) and externalizing (e.g.
delinquency/aggression) scores. This provided a measure of psychopathology for the
study. Internalizing and externalizing psychopathology scores were dichotomized
and coded (0,1) with adolescents coded as 0 for "normal" if they scored 0-70 (the
norm cutoff point) and adolescents scoring over 71+ (the clinical cutoff point) coded
as 1 for “clinically significant”. The validity and reliability of these widely used
instruments are well documented (Achenbach, 1991). Achenbach (1991) reported
that the mean 7-day test-retest reliabilities for the problem scales were .65 for 11- to
14-year-old adolescents and .83 for 15- to 18-year-old adolescents.
Generality of Violence. Generality of violence will be measured using the
Personal History Questionnaire (PHQ) (Riggs et al., 1990). Whereas generality of
violence refers to violence outside the marriage in the adult batterer literature,
generality of violence in the present study refers to violence occurring outside the
adolescent dating relationship. This is evaluated by assessing the respondents’
history of physical aggression not occurring exclusively in the context of a dating

45
relationship. Each respondent was asked to indicate how often he or she got into a
physical fight during elementary, junior high, and high school. Responses were
indicated on a 5-point scale from 0 = never to 4 = often. Items were summed to
obtain an overall score. Cronbach’s alpha will be separated for both males (.68) and
females (.70).
Contextual factors of dating violence
This construct will be measured using two instruments: the Reaction to
Physical Aggression scale, which assesses the respondents’ reaction to dating
violence, and the Reason for Physical Aggression scale, to assess the reasons a
respondent uses dating violence.
Reaction to Physical Aggression Measure. Reaction to physical aggression in
the dating relationship was measured by asking respondents to write a 1 for their first
choice and a 2 for their second choice from a list of possible responses to the
question, “How did you react to your partner’s physical aggression toward you?”
The list included the following items: fear, anger, emotionally hurt, thought it was
funny, wanted to get revenge, sad/depressed, ignored it, felt guilty, no effect at all,
wanted to get away, helplessness, afraid of not being loved, felt partner loved me,
shame, humiliation, and felt deserved.
Respondents could endorse up to four reactions to their violence from among
16 choices (including “other”). About 1/3 (n=368) of the respondents answered this
set of questions and gave 1007 reactions (average of 2.7 reactions per respondent)
for their violence. The top three answers were emotionally hurt (36%), anger (35%),

46
and fear (24%). The 16 reactions were then grouped into four clinical categories:
anger (including anger, wanted to get revenge), sadness (emotionally hurt,
sad/depressed, guilty, humiliation, felt I deserved it), fear (fear, afraid of not being
loved, felt partner loved me) and other (ignored it, no effect). The four responses
were dummy coded into three variables, each with 0 or 1 as values. Respondents
who selected other or self-defense, or who did not answer the question, would appear
as zeros in the anger and the jealousy variables.
When respondents are classified into more than one category, priority coding
will be assigned as follows: sadness, then anger, and then fear. Thus, anyone with
any of their original four options coded as sadness will be dummy coded as sadness.
Anyone who is not dummy coded as sadness but who has anger in any of their
original four options will be dummy coded as anger. Anyone who is not coded as
sad or angry in any of their original four options will be dummy coded as fear (if
they have a fear code). If they are coded as other or are missing on all four of the
original variables, they will be coded as zero in all dummy variables.
Reason for Physical Aggression Measure. Reason for physical aggression in
the dating relationship was measured by asking respondents to write a 1 for their first
choice and a 2 for their second choice from a list of possible responses to the
question, “What were the main reasons for your use of physical aggression in the
relationship?” The list included the following items: to show anger, was unable to
express self verbally, to feel more powerful, to get control over other person, to get
back at him/her, was really mad at someone/something else, in self- defense, to

47
punish person for wrong behavior, to prove love, to get attention, jealousy,
drug/alcohol, and other. Respondents could select up to two reasons for their
violence from among 13 choices (including “other”). About 1/3 (n =356) of the
respondents answered this set of questions and gave 640 reasons (average of 1.8
reasons per respondent) for their violence. The top three answers were to show
anger (33%), to get attention (20%), and anger at someone or something else (19%).
The 13 reasons were then grouped into three clinical categories: controlling
(including anger, feeling more powerful, getting control over the other, anger at
someone/thing else, punishing the person for wrongdoing, and drug/alcohol use);
attention seeking (unable to express self verbally, to prove love, to get attention,
jealousy); and other/self-defense. The three responses were then dummy coded into
two variables with 0 or 1 as values. Respondents who selected other or self-defense,
or who did not answer the question, would appear as zeros in the controlling and the
attention seeking variables. When respondents are classified into more than one
category, priority coding will be assigned as follows: controlling, then attention
seeking. Thus, anyone with either of their original two options coded as controlling
will be dummy coded as controlling. Anyone who is not dummy coded as
controlling, but who has attention seeking in either of their original two options, will
be dummy coded as attention seeking. If they are coded as other or are missing on
both of the original variables, they will be coded as zero in all dummy variables.
Relationship Satisfaction. Relationship satisfaction was measured using the
Relationship Assessment Scale (RAS) (Hendrick, 1988) to assess satisfaction in a

48
dating relationship. The RAS is a seven-item (e.g. “How well does your partner
meet your needs?”), Likert-type scale, with each item rated on a 5-point answer
continuum. Cronbach’s alphas were separated for both males .82 and females .85 in
the sample.
Relational Conflict. Relationship conflict was measured using the Conflictual
Relationship Scale (Braiker & Kelley, 1979) to assess conflict in relationships. The
five-item scale assesses the frequency of arguments and the seriousness of problems
in close relationships. For example, “How often do you and your partner argue with
one another?” Each item included a 5-point Likert-type scale, with the higher scores
indicating greater relationship conflict. Cronbach’s alphas were separated for both
males .73 and females .81 in the sample.
Control, Distal, and Proximal variables
The selection of distal and proximal risk factor questionnaires were based on
Riggs & O’Leary’s (1989) model, and were used to verify differences between the
perpetrator subtypes’ cluster groups. To describe this construct, this study used two
control variables, socioeconomic status and age, and four distal variables: witnessing
inter-parental aggression, experiencing parent-child physical aggression, acceptance
of violence, and school/community violence. The study also used four proximal
variables: alcohol/drug use, seriousness of relationship, and the number of dating
partners/ length of time dating.

49
Control variables
SES. Socioeconomic status was determined using the Hollingshead 2-factor
scale (1975), based on parents’ education and occupation, and coded into 3
categories: high, medium, and low socioeconomic status. SES was used as a
covariate in the study because low-SES has been found to be a risk factor for dating
violence (O’Keefe & Treister, 1998; Hotaling & Sugarman, 1986; Stets &
Henderson, 1991)
Age. Age was measured as the current age in years at the time of the survey
and ranged from 13 years to 20 years of age. Age was also used as a covariate
because the severity of dating violence has been found to increase with age
(O’Leary, Malone, & Tyree, 1994; Feld & Strauss, 1989).
Distal variables
Witnessing Inter-parental Aggression. The Physical Aggression subscale of
the Conflict Tactic Scale (Straus et al., 1979) is an 8-item self-report instrument used
to measure the amount of violence witnessed by a child, including husband-to-wife
violence and wife-to-husband violence. Items were rated on a seven-point Likert
scale ranging from 0 to 6, with higher scores indicating more violence witnessed by
child.
Respondents were asked to consider the worst year of their childhood and
indicate the frequency with which they witnessed their parents use violent tactics
toward the other during the year, using a seven-point Likert scale ranging from 0 =
None, 1 = Once, 2 = Twice, 3 = 3-5 times, 4 = 6-10 times, 5 = 11-20 times, and 6 =

50
More than 20 times. The measure included the following items: threw an object at the
other; pushed, grabbed, or shoved the other; slapped or spanked the other; kicked,
bit, or punched the other; hit or tried to hit the other with an object; beat up the other;
threatened the other with a knife or gun; and used a knife or gun against the other.
The violent acts assessed include items 1 through 8 noted above (item nine, forced
sexual activity, was excluded). The same eight violent behaviors discussed above
were addressed using a 7-point interval level response scale. Scores for the child
witness of mother-father and father-mother violence were summed. The CTS was
used by O’Keefe and Treister (1998) in an adolescent sample, where they found
witnessing inter-parental violence was associated with dating violence. Cronbach’s
alpha was separated in the sample for both males .93 and females .94.
Research has shown that the CTS had good psychometric properties. The
instrument is estimated to have relatively high internal consistency, high reliability,
and stable factor structures (Barling et al., 1987; Straus, 1979; Straus et al., 1980),
and has been used in numerous studies of dating aggression (Arias et al., 1987;
Henton et al., 1983; Sigelman et al., 1984; Stets and Pirog-Good, 1987). Straus and
Gelles (1990) reviewed findings that tested the internal consistency of the CTS. The
research showed alpha coefficients that were high for the physical
aggression/violence (r = .42 to .96). Although concurrent validity scores were
difficult to establish, there was “strong evidence of construct validity” (Straus &
Gelles, 1990, p. 72). A meta-analytic review (Archer, 1999) has shown that for
measuring aggression, the CTS is “better than the less commonly used alternatives”

51
(p. 1272). This is significant, because several instruments in this study use the CTS
to measure aggression within different situations: mother and father, parent and
child, and school/community member to school/community member.
Experiencing Parent-Child Physical Aggression. Experiencing physical
aggression from parents was assessed using the Physical Aggression subscale of the
Conflict Tactics Scale (Parent-Child Version) (Straus, 1979). Straus and Hamby
(1997) reported high construct validity for the CTS for parent-child violence.
Respondents report the number of incidents of mother-child and father-child physical
aggression experienced. Items are rated on a seven-point Likert scale ranging from 0
to 6, in which 0 = None, 1 = Once, 2 = Twice, 3 = 3-5 times, 4 = 6-10 times, 5 = 11-20
times, and 6 = more than 20 times. Respondents were asked to consider the worst
year of childhood and indicate the frequency that each parent used violent tactics
toward them during that year. The respondents were given eight items depicting nine
different types of violence: 1) threw something at me; 2) pushed, grabbed, or shoved
me; 3) slapped/spanked me; 4) kicked or hit me with a fist; 5) hit or tried to hit me
with something, for example, a belt; 6) beat me up; 7) threatened me with a knife or
gun; and 8) used a knife or gun. Scores for mother-child and father-child violence
were summed. O’Keefe (1997) used the CTS to measure parent-child physical
aggression in her adolescent dating violence study. The Cronbach’s alpha for
experiencing parent-child physical aggression for males was .87 and females .85.
Acceptance of Violence. Acceptance of violence refers to the level of which a
person approves of dating violence. The Justification of Violence Scale was adapted

52
from a scale developed by Margolin and Foo (1992) to assess attitudes toward dating
violence.
Respondents were told, “Some people think it is all right to hit or slap under
certain conditions,” and after this sentence they were asked, “After each of the
following statements indicate whether or not you think it is OK for couples to slap or
hit one another under the following situations.” Respondents were asked to select
from four options: never all right, sometimes all right, usually all right, and all right.
Then, they were asked whether they believed violence was justifiable in response to
the following nine situations: He or she (a) threatens to break up with the other, (b) is
drunk or drugged and acting crazy, (c) hits first in an argument, (d) finds out the
other is cheating on her or him, (e) calls the other nasty names, (h) refuses to let the
other go out for an evening with friends, and (i) is there any other time it is OK. To
further assess acceptance of violence, the Justification of Violence Scale was used.
This is an interval level of measurement using a 4-point Likert scale, which assesses
a respondent’s level of acceptance for a boy to hit or slap his girlfriend, and for a girl
to hit or slap her boyfriend, under nine situations. Scores were summed to indicate
justification of boy-to-girl violence and girl-to-boy violence. Both measures will be
used in this study. Scores for Justification of Girl-to-Boy Violence Scale and
Justification of Girl-to-Boy Violence Scale were summed. O’Keefe (1997) used
Justification of Violence Scale in her study on dating violence and found a Cronbach
alpha of .81 calculated for Justification of Boy-to-Girl violence and .87 for
Justification of Boy-to-Girl Violence scale, showing high reliability for an adolescent

53
sample. The sample was separated by gender to calculate Cronbach’s alpha for the
study, resulting in boys .86 and girls .88.
School/Community Violence. A modified version of the Child Conflict
Tactics Scale was used for an 11-item self-report neighborhood scale, and a 9-item
self-report school scale, to measure exposure to community violence (Straus, 1979).
The Modified CTS was used as an instrument to measure the type and frequency of
violence witnessed in their community during the previous year. Respondents were
asked, “Have you ever seen acts of aggression/violence in your neighborhood?” and
“Have you ever seen acts of aggression/violence in your school?” Respondents were
asked to select either “Yes” or “No” for each item. Respondents that answered “Yes”
were asked to select from various types of violence witnessed—pushing, grabbing,
shoving, kicking, hitting with fist, hitting with object, beating up, threatening with
knife/gun, stabbing, shooting, burglary, vandalism, rape, and others—and were asked
to indicate the frequency of these as being witnessed daily, weekly, monthly, twice a
year, or once a year. Burglary and rape were excluded from the examples of violence
at school, leaving only nine items for that question. The school/community violence
scale used a 6-point Likert-type scale coded 0 = never and 5 = daily. The total
exposure to school/community violence score was constructed by adding exposure to
school violence and exposure to community violence total scores. Cronbach’s alpha
for school/community violence exposure for males was .93, and for females .92.
Several other pre-adolescent and adolescent dating violence studies used a similar

54
scale to measure school/community violence (Thompson & Massat, 2005; O’Keefe,
1997).
Proximal variables
Riggs & O’Leary (1989) define proximal variables as being closer to the
actual occurrence of dating violence. This study used four variables to describe this
construct: alcohol/drug use, seriousness of relationship, the number of dating
partners, and length of time dating.
Alcohol/Drug Use. The questions used in this research to assess the
frequency of alcohol/drug use were based on two studies (Cate, Henton, Koval,
Christopher, & Lloyd, 1982; Foo & Margolin, 1995). Alcohol and drug consumption
was measured in terms of usage patterns during the previous year. Respondents were
asked three different questions: “How often, if ever, do you drink alcohol? How
often, if ever, do you smoke marijuana? How often, if ever, do you do other drugs?”
The same five answers were provided for each question: almost daily, almost
weekly, about one a month, about once every 6 months, no more than once a year,
and never. Respondents were asked to select one of the items. The alcohol/drug use
scale used a 6-point Likert-type scale coded 0 = never and 5 = daily. The three items
were then summed to obtain a total score. Several other adolescent dating violence
studies used a similar scale to measure alcohol and substance use (Cate, Henton,
Koval, Christopher, & Lloyd, 1982; Foo & Marolin, 1995; O’Keefe, 1987).
Cronbach’s alpha for alcohol/drug use was .63 for males and .70 for females.

55
Seriousness of Relationship. Emotional commitment to partners is a key
factor to defining the seriousness of a relationship. Billingham (1987) created a
seriousness of relationship scale for the study of dating violence. Respondents were
asked, “How serious is or was the relationship?” Seven possible answers were
provided: causal date; someone you have dated often but are not emotionally
attached to; someone you are emotionally attached to but not in love with; someone
you are in love with; someone you are in love with and would like to marry, but have
not discussed marriage; someone you are in love with and have discussed marriage,
but made no plans; and someone you are in love with and have made plans to marry.
Respondents were asked to select one of the items. The seriousness of relationship
ranged from 1 = casual dating/little emotional attachment, to 7 = someone you are
engaged to marry. Billingham’s seriousness of relationship scale is an ordinal level
scale of measurement (Roberts, BrintzenhofeSzoc, Zebrack, & Behar, 2003). Several
previous adult and adolescent studies have used seriousness of relationship as a
measure evaluating emotional commitment in dating relationships (Cate et al., 1982;
Henton, Cate, Koval, & Christopher, 1983; Sigelman et al., 1984; Billingham, 1987;
O’Keefe, 1997; O’Keefe and Treister, 1998).
Number of Dating Partners. Another important characteristic is the number
of prior dating relationships each partner has had. Specifically, respondents were
asked, “How many boyfriends (or girlfriends) have you had since you began
dating?” Respondents were asked to select one of four possible answers: only one, a
few, 5 to 10, and more than 10. Response codes ranged from 1 = one to 4 = more

56
than 10. This is a discrete (categorical) or nominal level variable because the value of
any one category is no different than the value of another (Roberts et al., 2003). The
total score for each category describing a number of dating partners was calculated
by adding all items for each category. The measure of the number of dating partners
was used in prior studies on dating violence in college and adolescent samples (Stets
& Pirog-Good, 1987; O’Keefe, 1997).
Length of Time Dating. Another key characteristic of dating relationships is
the length of time of each relationship. Specifically, respondents were asked, “How
long have you dated that person?” A space was provided for respondents to indicate
the number of months they dated a partner. This is also a discrete (categorical) or
nominal level variable with all categories being no different than the value of
another. The total scores for length of time dating were calculated by adding all
items to obtain an overall score. The measure of the length of time dating was also
used in several dating violence studies using college and adolescent samples (Stets &
Pirog-Good, 1987; O’Keefe, 1997).
In conclusion, the data were coded and prepared for computerized analysis
using SPSS-PC 12.0. Descriptive statistics and other appropriate statistical analyses
was ordered, including a) means and standard deviation; b) Pearson Product
Correlation, to yield information about the inter-correlation between participants and
clustering variables, as well as distal and proximal variables; c) cluster analysis, to
classify participants into three or four relatively homogeneous groups; and d)

57
MANCOVA’s, to compare the groups on multiple dependent variables from the
literature review while controlling for demographic variables.

58
Chapter 5: Results
This chapter outlines the statistical analysis of differences among typologies
in terms of distal and proximal dating violence variables. The explanation of the
results is guided by the four research questions and their corresponding hypotheses,
which were the objects of this study. Following the provision of descriptive statistics
– including means and standard deviation – the chapter presents Zero-order Pearson
Product Moment Correlations, Cluster analysis, and MANCOVA.
Demographic information
The final sample included 972 high school students, ages 13-20 years with n
= 402 males (41.4%) and n = 570 females (58.6%). Racial/ethnic breakdown among
students included n = 119 (12.3%) African American, n = 238 (24.5%) White, n =
517 (53.3%) Latino, n = 69 (7.1%) Asian/Pacific Islander, n = 9 (.9%) Native
American, and n = 18 (1.9%) other. Whites were over-represented in the suburban
areas, while Latino and African American students were over-represented in the
inner city areas. The socioeconomic status was composed of three groups: n = 183
(20.4%) upper SES, n = 244 (27.2%) middle SES, and n = 471 (52.4%) lower SES.
The sample indicates n = 388 (42.6%) perpetrators and n = 550 (56.6%) subjects
who were not perpetrators. There are n = 147 (36.6%) male perpetrators and n = 241
(42.3%) female perpetrators. Only n = 55 (5.6%) were perpetrators alone. On the
other hand, only n = 85 (8.7%) were victims alone. More commonly, subjects were
both victim and perpetrator n = 333 (34.2%), or neither victim nor perpetrator n =
464 (47.7%).

59
SES was calculated with The Two Factor Index of Social Status
(Hollingshead, 1975), which measured family SES on the basis of the education and
occupation for parents of the participants. Occupation was classified into one of nine
categories. Validity of the occupational scale was originally established by
comparing scores from the Hollingshead Index with scores of evaluations of
occupations and occupational groups from the National Opinion Research Center.
Education attainment was categorized as less than 7th grade, 8th through 9th grades,
10th through 11th grades, high school graduate, some college education, college
graduate, and some graduate school. The scale value for occupation (multiplied by a
factor weight of five) was added to the scale value for education (multiplied by a factor
weight of three). Higher scores represent higher social status. Tables 1 and 2 include
the distribution of gender, age, and SES for each typology in the study.

Table 1
Gender, Age, and SES of Each Adolescent Male Typology Group
Age SES Males

M SD M SD N %
Generally Violent 16.74 1.155 3.75 1.251 38 11
Situational Violent 17.06 .979 3.81 1.121 79 22.8
Non-violent 16.89 1.114 3.55 1.184 229 66.2

60
Table 2
Gender, Age, and SES of Each Adolescent Female Typology Group
Age SES Females

M SD M SD N %
Depressed Attention
Seeker
16.82 1.172 3.65 1.016 55 11.1
Hostile Controller 16.82 1.018 3.47 1.085 87 17.5
Undifferentiated
Controller
16.78 1.250 3.42 1.015 106 21.3
Non-violent 16.80 1.109 3.45 1.155 249 50.1

Cluster Analysis
The first step of the cluster analysis included Cronbach alphas for all the
measures used in the analysis. The Cronbach alphas for each measure were listed in
the previous methods section and all fell within an appropriate range for the
reliability of each measure. The second step of the cluster analysis was conducted to
answer the research question: do typologies exist in adolescent dating violence? Data
were analyzed separately for males and females to explore the typologies of dating
violence among adolescents, by gender. Sub-questions included: a) Do typology
subtypes of dating violence exist in either gender? b) Do the subtypes differ by
gender? c) What are the characteristics of each subtype of dating violence?
Characteristics for each subtype, by gender, were identified using MANOVA, in
which the variables used for clustering served as the dependent variables.

61
Cluster analysis is a procedure that empirically forms clusters or groups of
highly similar entities. This multivariate statistical procedure starts with a data set
containing information about cases and attempts to reorganize the cases into
relatively homogeneous groups. Cluster analysis was used in this study to develop a
typology or classification of adolescent perpetrators of dating violence. The literature
review led to the selection of variables used in clustering the cases: severity of
violence, generality of violence, internalizing behavioral problem, externalizing
behavioral problem, history of physical aggression, relationship satisfaction, reaction
to dating violence (three variables: sadness, anger, and fear), and reason for dating
violence (two variables: control and attention seeking). Recipient of violence was
also added for female cluster analysis. The following sub-steps were used to conduct
a parsimonious cluster analysis.
1. The clustering variables were checked for missing data. Between 18
and 34 cases were missing from five of the variables (approximately
3% of the data). Relationship satisfaction had 10% of its data missing
(n=102), because not all respondents had experience with
relationships. No replacements for missing data were conducted.
2. The file was split by gender and variables were standardized by
dividing values by their gender-specific range. This approach is
supported by Milligan and Cooper (1988), who found that dividing by
the range was superior to other methods of standardization.

62
3. Frequency distributions revealed that three variables (violence
inflicted, externalizing behavior problem, and internalizing behavior
problem scores) were not normally distributed, with many subjects
having no problems or having inflicted no violence on others.
Therefore, those variables were dichotomized for analysis. Reactions
to violence and reasons for violence were already dichotomous
variables.
4. Next, a bivariate correlation matrix was produced to examine for
possible redundancy. Most of the intercorrelations were low (less
than r=.30), with only a few correlations achieving a moderate level
of r=.58 (males) and r=.53 (females). Thus, all variables were retained
for clustering.
5. A nonhierarchical clustering method was used in this study.
Nonhierarchical clustering results in discrete clusters and requires the
researcher to specify a number of clusters to extract prior to
conducting the analysis. Nonhierarchical clustering is most often
implemented with the K-means algorithm (MacQueen, 1967). With
this method, initial cluster centers (values representing the average of
each cluster on each variable) were chosen at random by the program,
as shown in Table 3 for males and Table 4 for females. For males,
there was a failure to converge after 25 iterations. Minimum distance
between initial centers was 2.313. For females, convergence was

63
achieved at the twenty-fifth iteration, and the minimum distance
between initial centers was 2.221.
Tables 3 and 4 display the average distances between clusters for each
gender.

Table 3
Average Distance Between Adolescent Male Clusters
Generally Violent Situational Violent NV
1
2 1.62
3 1.98 1.37

Table 4
Average Distance Between Adolescent Female Clusters

Depressed
Attention
Seeker
Hostile Controller
Undifferentiated
Controller
NV
1
2 1.414
3 1.370 1.110
4 1.796 1.447 1.793

64
6. Once initial centers were chosen, the program assigned cases to the
cluster whose center was nearest. This is depicted in Table 5 for males
and Table 6 for females. Assigning the cases in this manner usually
changes the clusters’ centers, and thus objects were reassigned to
clusters, and the centers were updated again in an iterative manner.
This process continued until no cases changed their cluster
memberships. Table 7 shows the final cluster centers for males and
Table 8 shows the final cluster centers for females.

Table 5
Initial Cluster Centers for Males
Cluster Variable Clusters

Generally
Violent
Non-violent Situational Violent
Severity of Dating violence 1.0 0.0 1.0
Generality of Violence .11 1.0 .33
Relationship Satisfaction .85 1.15 .46
Conflict in Relationships .83 .33 1.22
Externalizing Psychopathology 1.0 0.0 1.0
Internalizing Psychopathology 1.0 0.0 0.0
Sadness reaction 0.0 0.0 1.0
Anger reaction 1.0 0.0 0.0
Fear reaction 0.0 0.0 0.0
Controlling reason 0.0 0.0 1.0
Attention seeking reason 1.0 0.0 0.0

65

Table 6
Initial Cluster Centers for Females
Cluster Variable Clusters

Depressed
Attention Seeker
Hostile
Controller
Undifferentiated
Controller
Non-
violent
Severity of Dating
violence
1.00 .00 1.00 .00
Generality of Violence 1.00 1.00 1.00 .00
Relationship
Satisfaction
.11 .33 .56 .00
Conflict in
Relationships
.64 .57 1.07 1.00
Externalizing
Psychopathology
1.05 1.05 .55 .25
Internalizing
Psychopathology
.00 .00 1.00 .00
Sadness reaction 1.00 .00 1.00 .00
Anger reaction .00 .00 1.00 1.00
Fear reaction .00 1.00 .00 .00
Controlling reason 1.00 .00 .00 .00
Attention seeking
reason
.00 1.00 1.00 .00

66

Table 7
Final Cluster Centers for Males
Cluster Variable Clusters

Generally
Violent
Non-violent Situational Violent
Severity of Dating violence .84 .14 .94
Generality of Violence .61 .32 .37
Relationship Satisfaction .92 .96 .91
Conflict in Relationships .77 .60 .77
Externalizing Psychopathology 1.0 .05 .00
Internalizing Psychopathology .61 .07 .03
Sadness reaction .74 .05 .82
Anger reaction .11 .03 .11
Fear reaction .03 .00 .01
Controlling reason .68 .03 .80
Attention seeking reason .24 .05 .13

67

Table 8
Final Cluster Centers for Females
Cluster Variable Clusters

Depressed
Attention Seeker
Hostile
Controller
Undifferentiated
Controller
Non-
violent
Severity of Dating
violence
.98 .91 .84 .03
Generality of Violence .93 .86 .92 .06
Relationship
Satisfaction
.21 .23 .21 .11
Conflict in
Relationships
.90 .76 .84 .93
Externalizing
Psychopathology
.77 .78 .76 .55
Internalizing
Psychopathology
.15 .11 .18 .04
Sadness reaction .16 .10 .14 .06
Anger reaction .69 .00 1.00 .00
Fear reaction .13 .40 .00 .00
Controlling reason .05 .06 .00 .01
Attention seeking
reason
.00 .64 .86 .00

68
7. The seventh sub-step of the cluster analysis involved interpreting
coefficients, as well as output of cluster identification for each case
for the number of clusters selected. Three clusters were selected a
priori based on theory (K-means), and then the analysis was re-run
with 1 more and 1 less cluster supported by the theory in order to
explore the quality of different solutions. As shown in Table 7, for
males, relationship satisfaction and fear did not discriminate between
clusters. Cluster 1 cases scored moderately high on severity of
violence, sadness, and controlling reasons for violence, and high on
generality of violence, conflict in relationships, externalizing and
internalizing behavior problems, anger, and attention seeking reasons
for violence. This is the generally violent typology, made up of 38
cases. Cluster 3 cases (n=79) scored high on severity of dating
violence, conflict in relationships, sadness, anger, and controlling
reasons for violence, making up the situational violent typology.
Cluster 2, the non-violent typology (n=229), scored low on all
variables. As shown in table 8, fear did not discriminate between
female clusters. Cluster 1 cases scored moderately high on severity of
violence, recipient of violence, relationship satisfaction, reacting to
violence out of sadness and fear, and using violence due to attention
seeking. Cluster 1 scored moderately high on conflict in relationships
and low on externalizing and internalizing behavior problems, as did

69
clusters 2 and 3. This is the depressed attention seeker violent
typology, made up of 55 cases. Cluster 2 also scored moderately high
on severity of violence, recipient of violence, reacting to violence
with anger and fear, and using violence due to controlling reasons.
This represented the hostile controller typology (n=87). Cluster 3
cases also scored moderately high on severity of violence, recipient of
violence, conflict in relationship, reacting to violence out of sadness,
and controlling reasons for violence. This cluster represents the
undifferentiated controller typology, made up of 106 cases. Cluster 4
is the non-violent typology (n=249), scoring the highest on
relationship satisfaction with no other significant scores. The
typologies between males and females were very different, in that no
generally violent females were depicted. Instead, variations of
situational violent typologies emerged, largely due to contextual
violence process variables, such as reactions to and reasons for
violence. The male situational violent group also scored high on
severity of dating violence, conflict in relationships, reacting to
violence with sadness and anger, and using controlling reasons for
violence. This is similar to the female hostile controller typology and
the female undifferentiated controller typology. Non-violent males
and females had almost identical compositions of typology, except

70
that females scored high on relationship satisfaction (whereas
relationship satisfaction did not distinguish typology among males).
8. The eighth sub-step involved describing clusters obtained in terms of
their demographic and clinical characteristics. A good cluster solution
resulted in clusters that were fairly homogeneous within cluster on
these variables. This appears to be the case, with no significant
differences between clusters on continuous demographics (as tested
by one-way Analysis of Variance (ANOVA) followed by post hoc
Scheffe tests (using harmonic means), or associations between cluster
and demographic variables (tested using contingency table analysis)
for age, grade, or SES. The three racial/ethnic groups with sufficient
sample size for conducting a contingency table analysis were African
American (n=109), non-Latino Caucasian (n=176), and Latino
(n=454). Among males, African American and Latino subjects were
significantly more likely to be associated with the situational violent
cluster, whereas non-Latino Caucasians were more likely to be
associated with the non-violent cluster (chi-square(4)=12.87, p=.012).
The results for females found no differences between clusters on all
demographic variables including age, ethnicity, grade, and SES. The
third step of the cluster analysis was to compare clustering variables
by cluster. A good cluster solution demonstrates discriminant validity,
consisting of significant differences between clusters on all or most of

71
the clustering variables. Multivariate Analysis of Variance
(MANOVA) was used to examine cluster validity for the three
continuous variables (generality of violence, relationship satisfaction,
and conflict in relationships). Contingency table analysis was used to
examine cluster validity for the dichotomous variables (internalizing
and externalizing behavior problems, severity of dating violence,
sadness, anger, fear, controlling, and attention seeking). These
analyses were done separately for each gender. Table 9 displays the
means of the continuous variables by cluster for males, and Table 10
displays them for females. The MANOVA for males was significant
(F(6,684)=16.05, p<.001). There was a significant difference between
clusters on generality of violence and conflict in relationships
(F(2,343)=23.21, p<.001 and F(2,343)=30.83, p<.001, respectively),
but relationship satisfaction did not differ significantly by group. The
Scheffe post hoc analysis using harmonic means revealed that the
generally violent cluster scored significantly higher on generality of
violence than did the other two clusters (p<.05), and the generally
violent and situational violent clusters scored significantly higher on
conflict in relationships than did the non-violent cluster (p<.05). The
MANOVA for females was significant (F(9,1479)=18.09, p<.001).
There was a significant difference between clusters on generality of
violence, relationship satisfaction, and conflict in relationships

72
(F(3,493)=9.99, p<.001, F(3,493)=13.51, p<.001, and
F(3,493)=56.04, p<.001, respectively). The Scheffe post hoc analysis
using harmonic means revealed that the non-violent clusters scored
significantly lower on generality of violence than did the other three
clusters (p<.05), and the hostile controller cluster scored significantly
higher than the generally violent cluster on relationship satisfaction
(p<.05). The Scheffe post hoc analysis revealed that the non-violent
cluster scored significantly higher on relationship satisfaction than did
the hostile controller cluster and undifferentiated controller clusters
(p<.05), and the non-violent cluster scored significantly lower on
relationship conflict than did the other three clusters (p<.05).

Table 9
Mean Scores by Cluster for Males
Generally Violent Situational Violent Non-violent
Male n=38 Male n=79 Male n=229
0
SD
0
SD
0
SD
Generality of Violence 5.45 2.54 3.29 2.13 2.91 2.05
Relationship
Satisfaction
23.92 5.34 23.72 5.74 24.93 6.07
Conflict in
Relationships
13.90 3.13 13.87 3.67 10.76 3.50

73
Table 10
Mean Scores by Cluster for Females
Depressed
Attention Seeker
Hostile
Controller
Undifferentiated
Controller
Non-violent
Female n=55 Female n=87 Female n=106
Female
n=249

0
SD
0
SD
0
SD
0
SD
Generality of
Violence
1.89 2.18 2.11 2.24 1.84 2.07 1.03 1.60
Relationship
Satisfaction
25.22 6.50 21.41 6.91 23.51 6.42 26.17 6.11
Conflict in
Relationships
15.38 4.60 15.59 4.32 15.11 3.80 10.93 3.38

Results of the contingency table analysis revealed that for males, the
generally violent and situational violent clusters were significantly associated with
severity of violence compared to the non-violent clusters (p<.001). The generally
violent males were significantly more likely to have elevated internalizing and
externalizing behavior problem score levels (p<.03). The generally violent and
situational violent males were significantly more likely to have sadness and anger
reactions to violence than the non-violent clusters (p<.010). There was no significant
association between cluster and fear reaction for males (p=.365 and p=.258,
respectively). Generally and situational violent clusters were significantly associated
with an controlling reason for violence (p<.001). Generally and situational violent
male clusters were also significantly associated with an attention seeking reason for
violence (p<.001).

74
The contingency table analysis found that for females, the depressed attention
seeker, hostile controller and undifferentiated controller clusters were more
significantly associated with severity of violence than were the non-violent clusters
(p<.001). The depressed attention seeker, hostile controller and undifferentiated
controller female clusters were also found to be more significantly associated with
being a recipient of violence than were the non-violent clusters (p<.001). The
depressed attention seeker females and undifferentiated controller females were
significantly more likely to have elevated internalizing behavioral problems (p<
.028), while only undifferentiated controller females were significantly more likely
to have externalizing behavior problem score levels compared to other groups
(p<.001). The depressed attention seeker females were significantly more likely to
have sadness reactions to violence than the non-violent clusters (p<.001). The
depressed attention seeker females and the hostile controller females were
significantly more likely to have angry reactions to violence than the non-violent
clusters (p<.001). The depressed attention seeker and hostile controller female
clusters were more likely to have a fear reaction to violence compared to other
groups (p<.004). Hostile controller and undifferentiated controller violent clusters
were significantly associated with a controlling reason for violence (p<.001).
Depressed Attention Seeker female clusters were significantly associated with an
attention seeking reason for violence (p<.001). Thus, strong discriminant validity
was established for the female adolescent perpetrator typology models.

75
Table 11
Zero-order Pearson Product Correlation for Overall Population
Variables 1 2 3 4 5 6 7
1. P-C -
2. Spousal .509** -
3. Accpt V .113** .096** -
4. Com/Sch .225** .214** .137** -
5. Etoh/drug .144** .090** .036 .182** -
6. Serio-Rel .051 .144** .025 .021 .-016 -
7. No.dtng .068* .058 .118** .198** .183** .083* -
8. Lgth dtg .021 .118** .043 .071 .-022 .535** .067
Note. * p < .05; ** p < .01

Inter-correlations
Zero-order Pearson Product Moment Correlations among all variables in the
study are shown in Table 11. After controlling for age and SES, the partial
correlations among all variables in the study are shown in Table 12. Regardless of
typology, witnessing inter-parental violence and parent-child violence are
moderately, positively correlated, along with the length of the dating relationship and
the seriousness of dating relationships (r = .522, p = .000; r = .530, p = .000,
respectively), as seen in the entire study sample as well as Table 11. Table 13 found
that a negative relationship existed among the generally violent males compared to
the overall population, after controlling for age and SES, between witnessing inter-
parental violence and the number of dating partners (r = -.539, p = .005). However, a
positive relationship was found between witnessing inter-parental violence and the

76
length of time dating (r = .503, p = .010). Variables in the generally violent males,
after controlling for age and SES, that were positively associated with one another
included acceptance of violence and the length of dating relationship (r = .409, p =
.042). Lastly, the number of dating partners was positively associated between
alcohol and drug use among the male non-violent group (r = .426, p = .001).

Table 12
Partial Correlation When Controlling Age and SES for Overall Population
Variables 1 2 3 4 5 6 7
1. P-C -
2. Spousal .522** -
3. Accpt V .140** .117** -
4. Com/Sch .215** .186** .101** -
5. Etoh/drug .150** .099** .060 .222** -
6. Serio-Rel .059 .154** .030 .020 .-013 -
7. No.dtng .037 .059 .134** .200** .230** .079* -
8. Lgth dtg .027 .117** .041 .044 .-029 .530** .051
Note. * p < .05; ** p < .01

77
Table 13
Partial Correlation When Controlling Age and SES for Male Generally Violent
Group
Variables 1 2 3 4 5 6 7
1. P-C -
2. Spousal .651** -
3. Accpt V .387 .198 -
4. Com/Sch .-075 .-133 .030 -
5. Etoh/drug .123 .-011 .-090 .-060 -
6. Serio-Rel .142 .215 .222 .383 .-237 -
7. No.dtng .-233 .-539** .-286 .203 .028 .-189 .
8. Lgth dtg .220 .503* .409* .281 .001 .578** .-314
Note. * p < .05; ** p < .01

Table 14
Partial Correlation When Controlling Age and SES for Male Situational Violent
Group

Variables 1 2 3 4 5 6 7
1. P-C -
2. Spousal .390** -
3. Accpt V .094 .170* -
4. Com/Sch .186* .044 .066 -
5. Etoh/drug .101 .-021 .017 .193* -
6. Serio-Rel .-005 .089 .076 .-040 .023 -
7. No.dtng .-030 .-069 .026 .029 .136 .173* -
8. Lgth dtg .061 .085 .-115 .-089 .-069 .502** .-030
Note. * p < .05; ** p < .01

78
Table 15
Partial Correlation When Controlling Age and SES for Male Non-Violent Group
Variables 1 2 3 4 5 6 7
1. P-C -
2. Spousal .404** -
3. Accpt V .-043 .-033 -
4. Com/Sch .116 .294* .-111 -
5. Etoh/drug .155 .-089 .101 .195 -
6. Serio-Rel .089 .-054 .-146 .027 .022 -
7. No.dtng .-124 .-127 .122 .276* .426** .027 -
8. Lgth dtg .171 .016 .-244 .118 .-207 .544** .-030
Note. * p < .05; ** p < .01

Tables 16 through 19 display the Pearson partial correlations for each
adolescent female perpetrator group controlling for age and SES: the depressed
attention seeker; hostile controller; undifferentiated controller; and the non-violent
comparison groups. Several differences from Table 11’s overall population emerge,
including a moderately positive relationship between witnessing inter-parental
violence and parent-child violence, and a similar relationship between the length of
dating relationships and the seriousness of dating relationships. For instance, positive
associations were found between parent-child violence and school/community
violence (r = .363, p = .032) among the female depressed attention seeker group. In
addition, positive correlations were found between witnessing inter-parental violence
and the following variables: school/community violence, alcohol and drug use, and the
length of time dating within the female depressed attention seeker group (r = .511, p =

79
.002; r = .343, p = .043, r = .424, p = .011, respectively). Another difference compared
to the overall sample among the female depressed attention seeker group was a
positive association between the number of dating partners and school/community
violence (r = .422, p = .012). The hostile controller group also differed from the overall
population pattern with a positive association between parent-child violence and the
following variables: school/community violence, and the number of dating partners (r
= .318, p = .010; and r = .292, p = .019, respectively). Witnessing inter-parental
violence was positively correlated with school/community violence within the female
hostile controller group (r = .335, p = .007). Alcohol and drug use was positively
associated with the number of dating partners among the female hostile controller
group (r = .312, p = .012).
Among the female undifferentiated controller group, a pattern developed where
the seriousness of the dating relationship was positively correlated with acceptance of
violence (r = .348, p = .003). Acceptance of violence was also positively associated
with the length of time dating (r = .446, p = .000) among the undifferentiated
controller group.

80
Table 16
Partial Correlation When Controlling Age and SES for Female Depressed Attention
Seeker Group
Variables 1 2 3 4 5 6 7
1. P-C -
2. Spousal .632** -
3. Accpt V .-044 .011 -
4. Com/Sch .363* .511** .092 -
5. Etoh/drug .156 .343* .143 .295 -
6. Serio-Rel .131 .079 .-132 .-009 .044 -
7. No.dtng .-043 .191 .287 .422* .-004 .-274 -
8. Lgth dtg .331 .424* .-043 .223 .183 .648** .-236
Note. * p < .05; ** p < .01

Table 17
Partial Correlation When Controlling Age and SES for Female Hostile Controller
Group
Variables 1 2 3 4 5 6 7
1. P-C -
2. Spousal .593** -
3. Accpt V .112 .047 -
4. Com/Sch .318* .335** .125 -
5. Etoh/drug .154 .178 .-177 .228 -
6. Serio-Rel .-154 .028 .-158 .-163 .-216 -
7. No.dtng .292* .148 .077 .208 .312* .-203 -
8. Lgth dtg .-165 .082 .033 .-065 .-187 .456** .-176
Note. * p < .05; ** p < .01

81
Table 18
Partial Correlation When Controlling Age and SES for Female Undifferentiated
Controller Group
Variables 1 2 3 4 5 6 7
1. P-C -
2. Spousal .394** -
3. Accpt V .139 .088 -
4. Com/Sch .097 .067 .287* -
5. Etoh/drug .157 .113 .-038 .-016 -
6. Serio-Rel .156 .241* .348** .247* .-058 -
7. No.dtng .065 .173 .127 .117 .071 .218 -
8. Lgth dtg .103 .065 .446** .246* .-050 .527** .290*
Note. * p < .05; ** p < .01

Table 19
Partial Correlation When Controlling Age and SES for Female Non-Violent Group
Variables 1 2 3 4 5 6 7
1. P-C -
2. Spousal .549** -
3. Accpt V .025 .045 -
4. Com/Sch .114 .067 .-066 -
5. Etoh/drug .023 .060 .-001 .182* -
6. Serio-Rel .112 .216** .-074 .066 .044 -
7. No.dtng .-025 .087 .008 .119* .160* .125 -
8. Lgth dtg .040 .063 .-003 .129 .078 .492** .140*

82
Multivariate analysis of covariance
Multivariate Analysis of Covariance using age and SES as covariates was
conducted to test whether the typologies possessed significantly different risk factors
after controlling for age and SES. Two separate MANCOVAs were conducted, one
for males and one for females, with adolescent dating violence risk factors as the
dependent variables. Descriptive information was provided on the observed and
adjusted means for the distal and proximal adolescent dating violence risk factors
and proximal risk factors for males and females in Tables 20 through 23. Using the
Wilks’s lambda criterion for both MANCOVAs, the null hypothesis is rejected for
Ho1. Tables 24 and 25 display results of the multivariate F indicated that distal and
proximal risk factors were significantly different based on their adolescent
perpetrator typology groups for both males (F(8,550)=10.004, p<.000;
F(8,510)=4.504, p<.000, respectively) and females (F(12,128)=2.657, p=.002;
F(12,117)=4.960, p=.001, respectively).
A Bonferroni adjustment was used to reject the null hypotheses for Ho2 and
Ho3 as well as all the remaining hypotheses. The generally violent and situational
violent male perpetrator groups were found to have higher scores than the non-
violent comparison group on parent-child violence, witnessing inter-parental
violence, acceptance of violence, and school/community violence (p = .000, p =
.000, p = .000, p = .000, respectively). The situational violent group was found to
have higher scores than the non-violent group on acceptance of violence (p=000) and
school/community violence (p=000). The generally violent group also had higher

83
scores than the situational violent group on parent-child violence (p=000) and
witnessing inter-parental violence (p=000). Further, evidence was found to reject the
null hypothesis for Ho4 that the generally violent and situational violent group would
differ from the non-violent group, but not for Ho5 where the generally violent would
differ from the situational violent group. The generally violent males were found to
have significantly higher scores than the non-violent comparison group on alcohol
and drug use (p = .008) and the number of dating partners (p = .000). The situational
violent group also had higher scores than the non-violent group on the number of
dating partners (p = .000).
Hypotheses 6 through 11 focused on the female typologies and their
differences regarding risk factors. The null hypothesis was also rejected for Ho6 for
the depressed attention seeker and the hostile controller groups. However, there was
no evidence to reject the null hypotheses for Ho7 and Ho8. The depressed attention
seeker female group was found to have higher scores than the non-violent
comparison group on school/community violence (p = .002), and the hostile
controller female group was found to have higher scores than the non-violent group
on parent-child violence (p=002). The null hypothesis was also rejected for Ho9 for
the female adolescent typology groups. There was evidence to reject the null
hypotheses for Ho10 for higher scores on the number of dating partners for the
hostile controller female group compared to the depressed attention seeker group;
however, there was no evidence for Ho11. The depressed attention seeker female
group was found to have higher scores on seriousness of relationship, the number of

84
dating partners, and the length of time dating (p = .001; p = .000; p = 039) than the
non-violent comparison group. The hostile controller group was also found to have
higher scores than the non-violent group on alcohol and drug use (p=001) and the
number of dating partners (p=000). The undifferentiated controller female groups
were found to have significantly higher scores than the non-violent comparison
group on alcohol and drug use (p = .008), the seriousness of the relationship (p =
.001), and the number of dating partners (p = .000).

Table 20
Descriptive Statistics of Distal Risk Factors and Adolescent Male Perpetrator
Typology
P-C Spousal Accpt Viol Sch/Com Viol
M SD AM M SD AM M SD AM M SD AM
Generally
vlnt
25.76 21.28 25.92 19.20 21.73 19.17 29.32 10.71 29.04 49.91 21.69 49.72
Situational
vlnt
11.28 10.85 11.26 06.68 11.19 06.79 23.57 06.78 23.63 32.55 17.25 32.82
Non-violent 13.16 12.66 13.15 08.44 14.50 08.13 27.73 09.08 27.69 42.79 19.39 42.09

Table 21
Descriptive Statistics of Proximal Risk Factors and Adolescent Male Perpetrator
Typology
ETOH/DRUGS Ser Rel No. Dat Part Lngth tm datg
M SD AM M SD AM M SD AM M SD AM
Generally vlnt 04.31 03.80 04.36 03.58 01.91 03.61 01.55 01.21 01.57 07.06 07.46 07.22
Situational vlnt 02.56 02.82 02.54 03.10 01.36 03.11 00.86 01.00 00.86 05.63 05.93 05.66
Non-violent 03.20 03.17 03.23 03.57 01.76 03.53 01.75 01.82 01.72 08.23 09.45 08.08

85

Table 22
Descriptive Statistics of Distal Risk Factors and Adolescent Female Perpetrator
Typology
P-C Spousal Accpt Viol Sch/Com Viol
M SD AM M SD AM M SD AM M SD AM
Depressed Attn
Skr
13.48 09.67 13.31 12.88 18.43 12.61 25.75 8.94 25.53 32.91 19.56 32.32
Hostile
Controller
14.23 15.31 14.22 09.85 16.40 09.84 24.39 07.61 24.42 29.57 19.03 29.60
Undiff
Controller
12.60 10.91 12.70 09.25 14.31 09.39 25.08 07.48 25.16 28.23 20.71 28.49
Non-violent 09.02 11.02 09.02 06.69 13.23 06.69 23.10 07.06 23.11 23.33 16.40 23.34

Table 23
Descriptive Statistics of Proximal Risk Factors and Adolescent Female Perpetrator
Typology
ETOH/DRUGS Ser Rel No. Dat Part Lngth tm datg
M SD AM M SD AM M SD AM M SD AM
Depressed
Attn Skr
02.45 03.00 02.56 04.65 01.80 04.65 01.30 00.99 01.29 15.87 12.78 15.81
Hostile
Controller
03.01 03.18 03.00 03.72 01.63 03.72 01.33 01.14 01.33 11.60 15.69 11.59
Undiff
Controller
02.92 03.01 02.91 04.38 01.87 04.37 01.21 00.98 01.21 12.60 11.21 12.59
Non-
violent
01.85 02.22 01.83 03.73 01.60 03.73 00.82 00.76 00.82 9.78 12.67 09.80

86

Table 24
Male Adolescent Perpetrator Typologies: Distal and Proximal Risk Factors of
Dating Violence, Controlling for Age and SES
Source λ F df p Sig. Post-Hoc Test
Distal .768 10.004 8,588 .000
Parent-Child Violence 18.885 2,296 .000
Generally Vlnt > Situational Vlnt
Generally Vlnt > Non-vlnt
Inter-parental Violence 12.248 2,296 .000
Generally Vlnt > Situational Vlnt
Generally Vlnt > Non-vlnt
Acceptance of Violence 11.658 2,296 .000
Generally Vlnt > Non-vlnt
Situational Vlnt > Non-vlnt
School/Com Violence 16.476 2,296 .000
Generally Vlnt > Non-vlnt
Situational Vlnt > Non-vlnt
Proximal .870 4.504 8,510 .000
Alcohol/drug use 4.960 2,257 .008 Generally Vlnt > Non-vlnt
Serious of relationship 2.598 2,257 .076
Number of dating partners 12.136 2,257 .000
Generally Vlnt > Non-vlnt
Situational Vlnt > Non-vlnt
Length of time dating 2.872 2,257 .058

87

Table 25
Female Adolescent Perpetrator Typologies: Distal and Proximal Risk Factors of
Dating Violence, Controlling for Age and SES
Source λ F df p Sig. Post-Hoc Test
Distal .928 2.657 12,1287 .002
Parent-Child Violence 5.145 3,430 .002 Hostile Controller > Non-vlnt
Inter-parental Violence 2.671 3,430 .047
Acceptance of Violence 2.565 3,430 .054
School/Com Violence 4.983 3,430 .002 Depressed Attn Skr > Non-vlnt
Proximal .874 4.424 12,1179 .000
Alcohol/drug use 4.960 3,394 .001
Hostile Controller > Non-vlnt
Undiff Controller > Non-vlnt
Serious of relationship 2.598 3,394 .001
Depressed Attn Skr > Non-vlnt
D. Attn Skr > Hostile Controller
Undiff Controller > Non-vlnt
Number of dating partners 12.136 3,394 .000
Depressed Attn Skr > Non-vlnt
Hostile Controller > Non-vlnt
Undiff Controller > Non-vlnt
Length of time dating 2.872 3,394 .039 Depressed Attn Skr > Non-vlnt

88
Chapter 6: Discussion
Research findings and implications for future research and professional
practices are discussed in this chapter. First, a summary of the significant findings
organized around the research questions for cluster analysis and hypotheses are
provided. Second, an explanation and interpretation of the results is provided as they
relate to original research questions and the hypotheses. Third, potential
contributions to the professional knowledge base as well as implications of the
findings as they relate to clinical social workers are discussed. Lastly, the chapter
outlines the limitations of the study along with their implications for the future
research.
Summary of Significant Findings
Cluster Analysis research question 1
Research Question 1 asked: Are there typologies in adolescent dating
violence for males and females? A 3-cluster solution emerged for males and a 4-
cluster solution emerged for females. Question 2 asked: Do the typologies differ for
males and females? The typologies did differ between males and females, with males
closely following adult batterer models. However, females were less predictable
compared to other adult batterer models, with no generally violent group emerging in
their analysis. Question 3 asked: What is the pattern of scores on the identified
clustered variables? The generally violent males scored higher on severity of dating
violence, generality of violence, internalizing and externalizing psychopathology,
relationship conflict, and reacting to violence with sadness and anger with

89
controlling and attention seeking reasons using violence. In contrast, the situational
violent males scored higher on severity of violence, reacted to violence with sadness
and anger, and used controlling and attention seeking rationale for violence.
The depressed attention seeker females scored higher on internalizing
psychopathology, reacted to violence with sadness, anger, and fear, used an attention
seeking rationale for violence, and had higher scores on relationship satisfaction
compared to some of the other groups. By comparison, the hostile controller females
reacted to physical aggression with anger and fear and used a controlling rationale
for violence, and had lower scores on relationship satisfaction than the non-violent
group. The undifferentiated controller group scored higher on both internalizing and
externalizing psychopathology and used a controlling rationale for violence.
Lastly, Question 4 asked: Are there differences among the typologies in
terms of risk factors, i.e. distal and proximal variables? This question led to 11
hypotheses for this study, each controlling for demographic variables:
(Ha1) found significant differences between male and female adolescent
perpetrator typologies on distal and proximal risk factors.
After controlling for age and SES, several of the null hypotheses about male
typology were rejected, such as:
(Ha2) The generally violent typologies were found to have higher scores on
distal risk factors compared to the non-violent groups.
(Ha3) The generally violent male typologies were found to have higher
scores on distal risk factors compared to the situational violent group.

90
(Ha4) The generally violent typologies were found to have higher proximal
risk factors compared to the non-violent typology group.
(Ha5) The generally violent males were found to have higher scores on the
proximal risk factors compared to the situational violent male perpetrators
group.
After controlling for age and SES, several of the null hypotheses about
female typology were rejected, such as:
(Ha6) The depressed attention seeker, the hostile controller, and the
undifferentiated controller violent typologies were found to have higher
scores on distal risk factors compared to the non-violent typology group.
(Ha7) The hostile controller, and the undifferentiated controller violent
typologies were found to have higher scores on distal risk factors compared
to the depressed attention seeker female perpetrators group.
(Ha8) There were no higher scores found for the hostile controller typologies
on distal risk factors compared to the undifferentiated controller violent
female perpetrators group.
(Ha9) The depressed attention seeker, the hostile controller, and the
undifferentiated controller violent typologies were found to have higher
scores on proximal risk factors compared to the non-violent typology group.
(Ha10) The hostile controller, and the undifferentiated controller violent
typologies were found to have higher scores on proximal risk factors
compared to the depressed attention seeker violent female perpetrators group.

91
(Ha11) There were no higher scores found for the hostile controller
typologies on proximal risk factors compared to the undifferentiated
controller violent female perpetrators group.
Statistical Results and Explanations
A priority of the study was to evaluate how adolescent perpetrator typologies
for males and females differed by proximal and distal risk factors. The study used
Riggs & O’Leary’s dating violence model to focus on specific risk factors that could
predict adolescent dating violence behavior (Riggs & O’Leary, 1989). Overall,
findings indicated that adolescent perpetrator typology groups differed by distal and
proximal risk factors. After revisiting the hypotheses and findings in the study,
several conclusions were drawn.
Descriptive statistics
With regard to several demographic characteristics for adolescent
perpetrators of dating violence, perpetration patterns differed within the adolescent
sample. As noted in the results chapter, 36% of male participants in the study, and
42.3% of female participants, described themselves as perpetrators. However, only
5.6% were perpetrators alone, and only 8.7% were solely victims. The majority of
adolescent perpetrators of dating violence were both victims and perpetrators at
34.2%. It is widely recognized that the majority of dating violence is reciprocal in
nature. Henton et al. (1983) and Gray and Foshee (1997) have found 70% of the
adolescents involved in violent relationships have reported mutual or reciprocal
violence.

92
Another common pattern in the adult typology literature is that there are
fewer subjects among typology groups with more severe characteristics (i.e.
generally violent, borderline/dysphoric) (Holtzworth-Munroe et. al, 2000). This was
also depicted in the results chapter with only 11% of the adolescent male sample
falling into the generally violent group, followed by 22.8% in the situational violent
group. This pattern was less clear in the adolescent female sample, with 11.1%
falling into the depressed attention seeker group, followed by 17.5% in the hostile
controller group, and 21.3% in the hostile controller group. The characteristics of
these typologies are very similar in terms of the severity and generality of violence.
The typologies differ on psychopathology and context of violence variables.
Interestingly, 34.2% of the adolescent sample is neither a victim nor a
perpetrator. This trend continues among the typology groups for both male and
female adolescent samples, with 66.2% of the males and 50.1% of the females
falling into the non-violent group.
Cluster analysis
The preliminary cluster provided a 3-cluster solution for adolescent males
and a 4-cluster solution for adolescent females. The adolescent male typology group
had some resemblance to Holzworth-Munroe, et al. (1994; 2000) and Johnson and
Leone (1995) batterer typologies for men. Three adolescent male groups emerged:
the generally violent, the situational violent, and the non-violent. A
borderline/dysphoric group did not emerge for this male adolescent sample as it did
for Holzworth-Munroes et al.’s adult batterer sample. It is possible that the

93
personality characteristics that are inherent in adult borderline/dysphoric batterers
had not fully developed in the male adolescent perpetrator sample.
The male adolescent typology groups can be described by the cluster
dimensions, which were statistically validated by the MANOVA analysis. The
generally violent males were found to have higher scores in severity of dating
violence, generality of violence, internalizing and externalizing psychopathology,
relationship conflict, reacted to physical aggression with sadness and anger, and
expressed controlling and attention seeking reasons for their own physical aggression
in the relationship. These findings are consistent with Holtzworth et al.’s (2000)
findings, except that adolescent generally violent males were found to have higher
scores on internalizing and externalizing psychopathology and reaction to violence,
as well as reasons for violence that were not measured in Holtzworth et al.’s study.
A reaction of anger in this study was described as feelings of anger and need
for revenge, while a reaction of sadness was described as feeling emotionally hurt,
sad/depressed, guilty, humiliated, and believing the other to be deserving the
aggression. A controlling rationale for violence in this study was described as
incorporating anger, feeling powerful, having need for control, displaced anger,
punishing, and alcohol and drug use. An attention seeking rationale for physical
aggression was described as an inability to express oneself verbally, misguided love,
need for attention, and jealousy.
The situational violent male group is different from the generally violent
male group in this study. This group closely resembles the Holtzworth et al. (2000)

94
family-only group as well as Johnson and Leone’s (1995) situational couples
violence group for adult male batterers. Situational couple violence occurs from a
specific conflict situation that escalates to violence. The situational violent male
group is descriptively similar to the generally violent group. However, the situational
violent adolescent male group was found to have moderately high scores in severity
of violence, with high scores in reacted to physical aggression out of sadness and
anger, and used both controlling and attention seeking rationale for violence against
their dating partners. However, similar to the Holtzworth et al. (2000) family only
adult male batterers, the situational violent males in this study were not found to
generalize violence to others or show signs or symptoms of psychopathology.
Female adolescent perpetrator typology differed from the adolescent male
perpetrator typologies in many ways. The most obvious was the 4 cluster solution for
females versus the 3 cluster solution for males. The next striking difference was that
a generally violent perpetrator group did not emerge for females. Instead, the groups
that did emerge closely resembled each other, but differed on psychopathology and
contextual factors of dating violence. Their names mirror these similarities and
differences: depressed attention seeker, hostile controller, and undifferentiated
controller groups emerged, as well as a non-violent comparison group. The three
groups showed no differences on the following cluster dimensions: severity of
violence, generality of violence, relationship conflict, and recipient of violence. The
depressed attention seeker female group was described as having higher scores on
internalizing psychopathology and reacted to dating violence with sadness, anger,

95
and fear. Fear in this study was defined as general fear, fear of not being loved, and
irrational love. The depressed attention seeker female group used an attention
seeking rationale for her physical aggression. This group scored higher on
relationship satisfaction than the hostile controller group.
The hostile controller female group also differed from the depressed attention
seeker female group in that no signs or symptoms of psychopathology were found.
However, like the depressed attention seeker female group, the hostile controller
group reacted to physical aggression with anger and fear, but unlike the depressed
attention seeker female group, they did not react to aggression out of sadness. The
hostile controller female used a controlling rationale for physical aggression in their
dating relationship. The hostile controller female group and the undifferentiated
controller group were found to have lower scores on relationship satisfaction than the
non-violent group.
The undifferentiated controller group also differed from the depressed
attention seeker and hostile controller females by scoring high on both internalizing
and externalizing psychopathology. Surprisingly, no statistically significant finding
occurred for reactions to physical aggression by the undifferentiated controller group
compared to the other groups. However, similar to the hostile controller female
group, the undifferentiated controller female group was found to use a controlling
rationale for their dating violence, differing from an attention seeking rationale used
in the depressed attention seeker female group.

96
Correlation statistics
The correlations in the study related to risk factors were moderate to
moderately high correlations. Witnessing inter-parental violence and parent-child
violence were moderately correlated along with the length of the dating relationship
and the seriousness of dating relationship, regardless of typology, for both males and
females. The typology groups for males and females differed from the overall sample
in many ways. For example, as witnessing inter-parental violence increased, the
number of dating partners decreased among the generally violent male group.
Bandura (1973) has argued that children growing up in abusive households learn that
physical aggression is a viable means for dealing with interpersonal conflict, thus
increasing the likelihood that the behavior will be re-enacted in later development. It
is possible that adolescent perpetrators learn to isolate their victims in order to
perpetrate violence against them without interference. This may explain why in this
study, as the risk of inter-parental violence increased, the number of dating partners
decreased.
Several other risk factors were found to be related among the generally
violent males. For instance, the history of witnessing inter-parental violence was
positively related to acceptance to violence, as well as to length of time dating. This
may suggest that the behavior and beliefs are learned in the home, leading to
unconscious or conscious strategies to prevent the victim from leaving the
relationship. Bowlby (1988) has stipulated that the “insecure attachment” which
develops between the parent-child dyad may lead to aggression in the relationship if

97
the perpetrator fears the possible loss of a partner. Therefore, perpetrators go to great
lengths to keep their victims in the relationship. The only risk factors associated with
the male non-violent group were an increase in the number of dating partners and an
increase in alcohol and drug use. It is possible that a decrease in the adolescent’s
inhibitions related to alcohol and drug use may lead to an increase in the number of
dating partners.
The female adolescent sample also differed from the overall sample on the
relationship between risk factors. As the risk of parent-child violence increased, so
did the risk for exposure to community violence among the female depressed
attention seeker group. Similarly, as the risk of witnessing inter-parental violence
increased in the female depressed attention seeker group, so did the risk of
school/community violence, alcohol and drug use, and the length of time dating. The
increase of the number of dating partners was also found to be related to an increase
in school/community violence in the female depressed attention seeker group.
The hostile controller females differed from the overall population in that an
increase in the risk of parent-child violence was related to an increased risk in
school/community violence, and the number of dating partners. Similarly, an
increase in the risk of witnessing inter-parental violence was related to an increase in
school/community violence, and an increase in alcohol and drug use was related to
an increase in the number of dating partners among the hostile controller female
group.

98
An increase in the seriousness of the dating relationship was associated with
an increase in the acceptance of violence in the undifferentiated controller females.
An increase in acceptance of violence was also related to an increase in the length of
time dating in the undifferentiated controller females.
It is possible that children learn dating violence in the home, but that such
violence gets reinforced in their community. This is consistent with Malik et al.’s
(1997) findings that exposure to school/community violence was associated with the
perpetration of dating violence. It is also possible that teens exposed to violence in
the home and to school/community violence may develop social skills deficits and
present themselves awkwardly to others, causing them to be excluded from positive
peer activities. Consequently, they may be left to befriend the outcasts in the
adolescent community. These groups may condone violence and alcohol and drug
use.
Alcohol and drugs have been strongly correlated with the intimate partner
violence in the adult literature, especially for the generally violent male group
(Holtzworth et al. (1994; 2000). It is also possible that the child learns to use alcohol
and drugs as a way of coping with both the perpetuation of violence as well as being
a victim of violence, and this behavior is reinforced by their social group. Therefore,
it seems plausible that they would re-enact violence and drug use in their dating
relationships. Alcohol and drug use may further prevent the victim from leaving the
relationship by allowing the person to tolerate prolonged suffering (e.g. through
numbing of feelings or senses, distorting perceptions, fostering denial, etc.).

99
Multivariate analysis of covariance
The study found significant differences between male and female adolescent
perpetrator typologies on distal and proximal risk factors (Ho1). This is not
surprising, given that risk factors from Riggs and O’Leary’s (1989) dating violence
model are expected to identify adolescent perpetrators of dating violence. The
generally violent and the situational violent male perpetrators scored significantly
higher on distal risk factors compared to the non-violent groups (Ho2). Specifically,
the generally violent male perpetrators were found to have higher scores than the
non-violent comparison group on parent-child violence, witnessing inter-parental
violence, acceptance of violence, and school/community violence. These findings are
consistent with Holzworth-Munroe et al.’s (2000) findings for the adult generally
violent perpetrator group. The situational violent group was found to have higher
scores on acceptance of violence and school/community violence. This may suggest
that the situational violent group believes there are times when violence is necessary
in the home, and that certain offenses warrant aggression. These beliefs may be
reinforced by exposure to violence in the school and community. The generally
violent males were also found to have higher scores on parent-child violence and
witnessing inter-parental violence compared to the situational violent group (Ho3).
This is consistent with Holtzworth-Munroe et al.’s (2000) findings that the generally
violent adult male batterers and borderline/dysphoric batterers had higher scores than
family-only adult male batterers. The family-only adult male batterers have
characteristics similar to the situational violent males in this study.

100
The generally violent and situational violent males scored significantly higher
on proximal risk factors than the non-violent typology group (Ho4). The generally
violent males were found to have higher scores on alcohol and drug use and the
number of dating partners than the non-violent comparison group, and the situational
violent males had higher scores than the non-violent group on the number of dating
partners. This may suggest that the generally violent and situational violent males are
more experienced at dating compared to the non-violent group. It is also possible that
these groups have a propensity for experimenting with alcohol and drugs.
The depressed attention seeker, the hostile controller, and undifferentiated
controller violent female typologies scored higher on distal risk factors than the non-
violent typology group (Ho6). Specifically, the study found that the depressed
attention seeker female group had higher scores than the non-violent comparison
group on school/community violence. Internalizing psychopathology is a
characteristic of the depressed attention seeker group. Consequently, it is possible
that these internalizing characteristics lead to ostracizing by peers. Therefore,
depressed attention seeker females may be left to befriend other social outcasts, who
may be more tolerant of actions such as school/community violence. The hostile
controller female group also had higher scores than the non-violent group on parent-
child violence. The hostile controller female uses a controlling rationale for engaging
in dating violence. It is possible that using violence for controlling reasons is a
learned behavior modeled after parent-child violence in the home. This would also
explain why the hostile controller female reacts to physical aggression with anger

101
and fear. The hostile controller female is torn between two learned ways of reacting
to violence: reacting to aggression by another with the anger and rage of the
perpetrator, and reacting out of the fear of being victimized.
The depressed attention seeker, the hostile controller, and the undifferentiated
controller violent typologies were found to have higher scores on proximal risk
factors than the non-violent typology group (Ho9). Specifically, the depressed
attention seeker females were found to have higher scores on seriousness of
relationship, the number of dating partners, and the length of time dating compared
to the non-violent comparison group. The internalizing psychopathology of the
depressed attention seeker females may foster a dependency on their partners. This is
similar to Bowlby’s (1988) theory that “insecure attachment” may lead to
relationship violence if the perpetrator fears the loss of her partner. Therefore, this
may lead to feelings that the relationship is more serious than it actually is. It may
also foster a mindset that she cannot do without a relationship, thus increasing the
number of dating relationships and the length of those relationships.
The hostile controller, and the undifferentiated controller violent typologies
scored higher on proximal risk factors than the depressed attention seeker violent
female perpetrators group (Ho10). The hostile controller group was found to have
higher scores than the non-violent group on alcohol and drug use and the number of
dating partners. The hostile controller female group was not characterized by
psychopathological variables. In other words, there was not internalizing or
externalizing psychopathology present in this female perpetrator group. It is possible

102
that alcohol and drugs may be used by the hostile controller females as a way of
pretending to act more mature and fit in with other kids who are considered “cool.”
In contrast, the undifferentiated controller female group is characterized by
internalizing and externalizing psychopathology, and they were found to have
significantly higher scores than the non-violent comparison group on alcohol and
drug use, seriousness of the relationship, and the number of dating partners. It is
possible that growing up in an abusive household may lead to an increase in
internalizing and externalizing psychopathology, as well as learning to deal with
interpersonal conflict with physical aggression. The combination of increased
psychopathology and the likelihood that learned physical aggression will be re-
enacted in a later dating relationship provide the perfect storm for dating violence to
occur. This may foster a fear of being alone, and not in a relationship. Therefore, the
perpetrator may become dependent on the relational partner and fearful that she may
leave. This would increase the number of dating partners and the length of those
relationships. Lastly, as suggested earlier, the victim and the perpetrator may use
alcohol and drugs as a maladaptive coping mechanism to maintain the abusive
relationship.
Implications for Social Work Practice
The findings of this study confirm that adolescent perpetrator typologies
differ by distal and proximal risk factors for males and females. Moreover, this study
shows that typologies differ from one another within their gender group as well as
between gender groups (e.g. males vs. females). First, these findings suggest the

103
importance of a thorough initial assessment upon intake of a perpetrator of dating
violence to establish what characteristics describe them, such as severity of violence,
generality of violence, and psychopathology to name a few.
Second, these findings suggest that a thorough assessment of distal and
proximal risk factors, as they relate to perpetrator characteristics, could facilitate
placement of adolescent perpetrators into typology groups, matched to appropriate
treatments. Adolescents who perpetrate violence on a dating partner most likely do
so undetected. The perpetration of dating violence is done in silence and usually
remains in silence, sealed off from society by years of learned behavior and shame.
Furthermore, adolescents who are receiving mental health services are probably
doing so for reasons other than the perpetration of dating violence, such as
depression, conduct disturbances, and/or alcohol and substance use. Most likely,
mental health professionals are not asking adolescents suffering from these disorders
if they are perpetrating violence against their dating partners, and it is unlikely that
the adolescents are volunteering this information. Mental health professionals must
explore the possibility of dating violence perpetration with adolescents who present
with typology characteristics (i.e., externalizing psychopathology) or associated risk
factors (i.e., alcohol and drug use).
Third, the results of this study provide the rationale for clinicians to design
hybrid interventions to improve the treatment of adolescent perpetrators of dating
violence by focusing on adolescent perpetrator typologies in the context of their
environment. For example, clinicians who are aware that a profile matches the

104
situational violence typology for males or the hostile controller typology for females
can provide specific targets for existing evidenced-based treatment interventions,
such as cognitive-behavioral therapy for hostile attributions, directed at youth who
perpetrate violence as a reaction to their environment.
However, clinicians must be wary about only targeting one type of treatment
as a panacea for treating adolescent perpetrator of dating violence. Instead, it is
recommended that clinicians use several sources of information to assess and treat
these teens. For example, it is possible that the adolescent may suffer from several
deficits in their skill development, such as conflict resolution skills, adaptive coping
skills, communication skills, social skills, and problem solving skills to name a few.
Therefore, it would be important to include skills development in the treatment of the
adolescents. Other types of treatment may also be needed such as substance abuse
treatment, psycho-education on the history and etiology of dating violence, and
possibly family therapy. More extreme typologies, such as generally violent males,
may require a more intensive treatment protocol that provides them more frequent
contact with mental health professionals and increased structure.
Fourth, adolescent perpetrators of dating violence can be a focus for primary
and secondary prevention efforts by increasing the public’s awareness and providing
existing community agencies with the resources to provide prevention interventions
in the local community, including families, schools, churches, businesses, and other
institutions. Such prevention efforts might further reduce the risk and costs
associated with adolescent perpetrators of dating violence.

105
Limitations of the Study and Implications for Future Research
The study has several areas that can serve as a springboard for future
research. First, the data collected for this study was largely self-reported. Therefore,
the accuracy of the data is reliant on the honesty of the adolescent participants in the
study. Consequently, the study’s conclusions are limited by the participant’s
perceptions and knowledge. For example, the youth self report that measured
psychopathology in this study was restricted because the self reports were not
corroborated by responses from parent or teacher questionnaires. This limitation
resulted from time constraints and confidentiality issues. Future studies should
include a multi-pronged approach including gathering data from the child, parent,
and teacher to provide more comprehensive and accurate information.
Second, this was a cross-section study and was not designed to assess
adolescent typology groups and how they might change over time. Moreover, the
changes in the adolescent typology groups into adulthood could not be observed.
Thus, it is difficult to interpret the long-term effects of these adolescent typology
groups within these study limitations. Therefore, future research is needed to
systematically observe ongoing changes in typology groups over time using a
longitudinal design to improve understanding of perpetrator typologies throughout
the life span.
A third limitation of the study is its inability to generalize the distal and
proximal risk factors as dependent variables for the adolescent perpetrator
typologies. The study samples for adolescent perpetrator typologies were recruited

106
through purposive sampling from six high schools in the Los Angeles area. As a
result, the use of purposive sampling within the study has an inherent sampling bias.
Nevertheless, this research is very meaningful in that no other study has been
conducted to date on adolescent perpetrator typologies for males and females and the
differences among them. However, a randomized research design is needed for
future adolescent perpetrator typology research.
Fourth, there is always a possibility of spurious variables that could
potentially confound the findings of this study, such as coping strategies, sexual
abuse, attachment style, personality traits, and cultural factors. Therefore, future
studies should examine a more diverse selection of potential confounders, especially
for the female adolescent perpetrator typology groups.
Fifth, every study must contend with statistical trade off and choices in order
to ensure the best possible conclusions related to data. The data used in this study
violated an assumption of MANCOVA (homogeneity of dispersion matrices, as
assessed by the Box M test), and an assumption of ANOVA (homogeneity of
variance, as assessed by Levene’s test). Reasons for these violations may be due, in
part, to testing group differences for groups of radically different sizes (unequal N).
It is natural to expect unequal group sizes in a study of this nature, wherein the
majority of the sample studied can be expected to be non-violent, leaving only a
small minority of the sample to display the various typologies of violence.
Because age and socioeconomic factors were found in the literature to be
associated with youth violence, these two variables were co-varied throughout the

107
analysis, whether or not they met the statistical assumptions to be covariates. In fact,
almost all of the results demonstrated significant F ratios for these covariates,
although the adjustments of group means were relatively small. In analyzing
covariance, it is not possible to conduct typical post-hoc analyses of choice for
groups with unequal N (for example, the Scheffe test). Instead, the conservative
Bonferroni correction factor was used to control for inflation of family wise error
when conducting multiple comparisons.
While other statistical decisions could have been made (e.g. data
transformations to attempt to achieve homogeneity), the meaning of the results
would have been less accessible. Considering the intrinsic nature of the data, this
study used the best statistical options available.
Conclusion
Despite these limitations, this study provides meaningful new knowledge
about adolescent perpetrator typologies for males and females and the differences
among them. In particular, the evaluation of how adolescent perpetrator typologies
for males and females differed by distal and proximal risk factors is unique, such as
generally violent males vs. situational violent males or depressed attention seeker
females vs. hostile controller females. This study also examined descriptive
differences between adolescent male perpetrators and female adolescent perpetrators,
such as the fact that no generally violent group emerged from cluster analysis for
adolescent female perpetrators. Thus, this work will help to expand the knowledge
base for understanding the perpetration of violence and how to treat it. Moreover,

108
this study provides a link between adolescent male perpetrator typologies and adult
male batterer typology models, suggesting that interventions early in a child’s
development may be necessary to effectively prevent intimate partner violence. Once
more, this developmental connection between adolescent perpetrator typologies and
adult male batterer typologies suggests that a similar developmental link may exist
between female adolescent perpetrators and adult female perpetrators of intimate
partner violence. Future research is needed to evaluate the psychopathology and
contextual violence across the female perpetrator’s life span, as well as to explore
alternative risk factors for females.

109
References
Achenbach, T. M. (1991). Manual for the Youth Self-Report and 1991 Profile.
Burlington, VT: University of Vermont Department of Psychiatry.

Ainsworth, M.S. (1989). Attachments beyond infancy. American Psychologist, 44,
709-716.

Archer, J. (1990). Assessment of the reliability of the conflict tactics scales: A meta-
analytic model. Journal of Interpersonal Violence, 14, 12, 1263-1289.

Arias, I , Samios, M., and O'Leary, K.D. (1987). Prevalence and correlates of
physical aggression during courtship. Journal of Interpersonal Violence, 2, 1,
82-90.

Babcock, J., Jacobson, N. S., Gottman, J. M., & Yerington, T. P. (2000). Attachment,
emotional regulation, and the function of marital violence: Differences
between secure, preoccupied, and dismissing violent and nonviolent
husbands. Journal of Family Violence, 15, 391-409.

Babcock, J., Miller, S.A., & Siard, C. (2003). Toward a typology of abusive women:
Differences between partner-only and generally violent women in the use of
violence. Psychology of Women Quarterly, 27, 153-161.

Bandura, A., Ross, D. & Ross, S. A. (1961). Transmission of aggression through
imitation of aggressive models. Journal of Abnormal and Social Psychology,
63, 575-582

Bandura, A. (1969). Principles of behavior modification. New York: Holt, Rinehart
& Winston.

Bandura, A. (1973). Aggression: A social learning analysis. Englewood Cliffs, NJ:
Prentice-Hall.

Bandyard, V.L., Cross, C., and Modecki, K. (2006). Interpersonal violence in
adolescence: Ecological correlates of self-reported perpetration. Journal of
Interpersonal Violence, 21, 10, 1314-1332.

Barling, J., O'Leary, K.D., Jouriles, E.N., Vivian, D., & MacEwen, K.E. (1987).
Factor similarity of the conflict tactics scales across samples, spouses, and
sites: Issues and implications, Journal of Family Violence, 2, 37-53.

110
Bartholemew, K., & Horwitz, L.M. (1991). Attachment style among young adults: A
test of a four category model. Journal of Personality and Social Psychology,
61, 226-224.

Billingham, R.E. (1987). Courtship violence: The patterns of conflict resolution
strategies across seven levels of emotional commitment. Family Relations,
36, 283-289.

Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human
development. New York: Basic Books.

Braiker, H.B., & Kelley, H.H. (1979). Conflict in the development of close
relationships. In R.L. Burgess & T.L. Huston (Eds.), Social exchange in
developing relationships (pp. 135-168). New York: Academic Press.

Bussey, K., & Bandura, A. (1984). Influence of gender constancy and social power
on sex-linked modeling. Journal of Personality and Social Psychology, 47,
1292-1302.

Caesar, P.L. (1986). Men who batter: A heterogeneous group. In L.K. Hamberger
(Chair), The male batterer: Characteristics of a heterogeneous population.
Symposium conducted at the 94
th
Annual Convention of the American
Psychological Association, Washington, DC.

Carver, K., Joyner, K., & Udry, J. R. (2003). National estimates on adolescent
romantic relationships. In P. Florsheim (Ed.), Adolescent romantic relations
and sexual behavior (pp. 23–56). Mahwah, NJ: Erlbaum.

Cascardi, M., Avery-Leaf, S., O'Leary, K. D., & Smith-Slep, A. M. (1999). Factor
structure and convergent validity of the conflict tactics scale in high school
students. Psychological Assessment, 11, 546-555.

Cate, R., Henton, J., Koval, J., Christopher, R., & Lloyd, S. (1982). Premarital
abuse: A social psychological perspective. Journal of Marriage and the
Family, 3(1), 71-90.

Centers for Disease Control and Prevention. (2000). Youth risk behavior
surveillance-United States, 1999. Morbidity and Mortality Weekly Report, 49,
1-96.

Chase, K. A., O'Leary, K. D., & Heyman, R. E. (2001). Categorizing partner-violent
men within the reactive-proactive typology model. Journal of Consulting and
Clinical Psychology, 69, 567-572.

111
Coker, A. L., McKeown, R. E., Sanderson, M., Davis, K. E., Valois, R. F., &
Huebner, E. S. (2000). Severe dating violence and quality of life among
South Carolina high school students. American Journal of Preventive
Medicine, 19, 220-227.

Corvo, K., & Carpenter, E.H. (2000). Effects of parental substance abuse on current
levels of domestic violence: A possible elaboration of intergenerational
transmission processes. Journal of Family Violence, 15, 123-135.

Dodge, K.A., Lochman, J.E., Harnish, J.D., Bates, J.E., & Pettit, G.S. (1997).
Reactive and proactive aggression in school children and psychiatrically
impaired chronically assaultive youth. Journal of Abnormal Psychology, 106,
37-51.

Dutton, D.G., & Golant, S.K. (1995). The batterer: A psychological profile. New
York: Basic Books.

Dutton, D.G., Nicholls, T.L., & Spidel, A. (2005). Female perpetrators of intimate
abuse. Journal of Offender Rehabilitation, 41(4), 1-31.

Dutton, D. G., Saunders, K., Starzomski, A., & Bartholomew, K. (1994). Intimacy-
anger and insecure attachment as precursors of abuse in intimate
relationships. Journal of Applied Social Psychology, 24, 1367-1386.

Elbow, M. (1997). Theoretical considerations of violent marriages. Social Casework,
58, 515-526.

Faulk, M. (1974). Men who assault their wives. Medicine, Science, and the Law, pp.
180-183.

Feeney, J.A., & Noller, P. (1990). Attachment style as a predictor of adult romantic
relationships. Journal of Personality and Social Psychology, 58, 281-291.

Feeney, J.A., & Noller, P. (1992). Attachment style and romantic love: Relationship
dissolution. Australian Journal of Psychology, 44, 69-74.

Feld, S.L., & Strauss, M.A. (1989). Criminology, 27, 141-161.

Foo, L. & Margolin, G. (1995). A multivariate investigation of dating aggression.
Journal of Family Violence, 10, 4, 351-377.

Gandolf, E.W. (1988). Who are those guys? Toward a behavioral typology of
batterers. Violence and Victims, 3, 187-203.

112
Gottman, J.M., Jacobson, N.S., Rushe, R.H., Shortt, J.W., Babcock, J., La Taillade,
J.J., & Waltz, J. (1995). The relationship between heart rate reactivity,
emotionally aggressive behavior, and general violence in batterers. Journal of
Family Psychology, 9, 227-248.

Gray, H.M., & Foshee, V. (1997). Adolescent dating violence: differences between
one-sided and mutually violent profiles. Journal of Interpersonal Violence,
12, 1, 126-142.

Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment
process. Journal of Personality and Social Psychology, 52, 511-524.

Hendrick, S. (1998). A generic measure of relationship satisfaction. Journal of
Marriage and the Family, 40, 92-98.

Henning, K., Jones, A., & Holdford, R. (2003). Treatment needs of women arrested
for domestic violence: A comparison with male offenders. Journal of
Interpersonal Violence, 18, 839-856.

Henry, D.B., Tolan, P.H., & Gorman-Smith, D. (2005). Cluster analysis in family
psychology research. Journal of Family Psychology, 19, 1, 121–132.

Henton, J., Cate, R., Koval, J., Lloyd, S., & Christopher, S. (1983). Romance and
violence in dating relationships. Journal of Family Issues, 4, 467-482.

Hershorn, M., & Rosenbaum, A. (1991). Over- vs. undercontrolled hostility:
Application of the construct to the classification of maritally violent men.
Violence and Victims, 6, 151-158.

Hickman, L.J., Jaycox, L.H., & Aronoff, J.A. (2004). Dating violence among
adolesents: prevalence, gender distribution, and prevention program
effectiveness. Trauma, Violence, & Abuse, 5 (2), 123-142.

Hicks, D.J. (1965). Imitation and retention of film-mediated aggressive peer and
adult models. Journal of Personality and Social Psychology, 2, 97-100.

Hollingshead, A. (1975). Four-factor index of social position. New Haven, CT:
Author.

Holtzworth-Munroe, A., Meehan, J. C., Herron, K., Rehman, U., & Stuart, G. L.
(2000). Testing the Holtzworth-Munroe and Stuart (1994) batterer typology.
Journal of Consulting and Clinical Psychology, 68, 1000-1019.

113
Holzworth-Munroe, A., & Struart, G. (1994). Typologies of male batterers: Three
subtypes and the differences among them. Psychological Bulletin, 116, 476-
497.

Holtzworth-Munroe, A., Stuart, G. L., & Hutchinson, G. (1997). Violent versus
nonviolent husbands: Differences in attachment patterns, dependency, and
jealousy. Journal of Family Psychology, 11, 314-331.

Hotaling, G. T. & Sugarman, D. B. (1986). An analysis of risk markers in husband to
wife violence: The current state of knowledge. Violence & Victims, 1, 101-
124.

Hulley, S.B., & Cummings, S.R., (1988). Designing clinical research. Baltimore:
Williams and Wilkins.

Jacobson, N., & Gottman, J. (1998). When men batter women. New York: Simon &
Schuster.

Janoff-Bulman, R. (1992). Shattered assumptions: Towards a new psychology of
trauma. New York: Free Press.

Johnson, M. (1995). Patriarchal terrorism and common couple violence: Two forms
of violence against women. Journal of Marriage and the Family, 57, 283-
294.

Johnson, M., & Leone, J.M. (1995). The differential effects of intimate terrorism and
situational couples violence: Findings from the National Violence Against
Women survey. Journal of Family Issues, 26, 322-349.

Lewis, S., & Fremouw, W. (2001). Dating violence: A critical review of the
literature. Clinical Psychology Review, 21, 105-127

Luthra, R., & Gidycz, C. A. (2006). Dating violence among college men and women:
Evaluation of a theoretical model. Journal of Interpersonal Violence, 21,
717-731.

MacEwen, K.E. (1994). Refining the intergenerational transmission hypothesis.
Journal of Interpersonal Violence, 9, 350-365.

MacQueen, J.B. (1967). Some methods for classification and analysis of multivariate
observations. In L.M. Le Cam & J. Neyman, (Eds.), Proceeding of the 5
th

Berkeley Symposium on Mathematical Statistics, (Vol. 1 pp. 281-297).
Berkeley: University of California Press.

114
Malik, S., Sorenson, S., & Aneshensel, C. (1997). Community and dating violence
among adolescents: Perpetration and victimization. Journal of Adolescent
Health, 21, 291-302.

Margolin, G., & Foo, L. (1992). The Attitudes About Dating Scale. Unpublished
measure at the University of Southern California.

Mayseless, O. (1991). Adult attachment patterns and courtship violence. Family
Relations, 40, 21-28.

McCloskey, L., Fegueredo, A., & Koss, M. (1995). The effects of systemic family
violence on children's mental health. Child Development, 66(5), 1239-1261.

Milligan, G.W., & Cooper, M.C. (1988). A study of standardization of variables in
cluster analysis. Journal of Classification, 5, 182-204.

Murphy, C. M., Meyer, S., & O'Leary, K. D. (1994). Dependency characteristics of
partner assaultive men. Journal of Abnormal Psychology, 103, 729-733.

National Center for Injury Prevention and Control. (2003). Costs of Intimate Partner
Violence Against Women in the United States, Atlanta (GA): Centers for
Disease Control and Prevention.

Neidig, P. H., Collins, B. S., & Friedman, D. H. (1986). Attitudinal characteristics of
males who have engaged in spouse abuse. Journal of Family Violence, 1,
223-233.

O'Keefe, M. (1997). Predictors of dating violence among high school students.
Journal of Interpersonal Violence, 12, 546-568.

O'Keefe, M., & Treister, L. (1998). Victims of dating violence among high school
students: Are the predictors different for males and females? Violence
Against Women, 4, 195-223.

O'Leary, K.D., Malone, J., & Tyree, A. (1994). Physical aggression in early
marriage: Prerelationship and relationship violence: A model of vulnerability
to victimization. Journal of Consulting and Clinical Psychology, 62, 594-
602.

O'Leary, K. D. & Slep, A. M. S. (2003). A dyadic longitudinal model of adolescent
dating aggression. Journal of Clinical Child and Adolescent Psychology, 32,
314-327.

115
Pistole, M.C., & Tarrant, N. (1993). Attachment style and aggression in male
batterers. Family Therapy, 20, 165-173.

Prince, J., & Arias, I. (1994). The role of perceived control and the desirability of
control among abusive and nonabusive husbands. American Journal of
Family Therapy, 22 (2), 126-134.

Ray, A. L., & Gold, S. R. (1996). Gender roles, aggression, and alcohol use in dating
relationships. Journal of Sex Research, 33, 47-55.

Riggs, D. S., & O'Leary, K. D. (1989). The development of a model of courtship
aggression. In M. A. Pirog-Good & J. E. Stets (Eds.), Violence in dating
relationships: Emerging social issues (pp. 53–71). New York: Praeger.

Riggs, D.S., O'Leary, K.D., & Breslin, F.C. (1990). Multiple correlates of physical
aggression in dating couples. Journal of Interpersonal Violence, 5, 530-548.

Roberts, C., BrintzenhofeSzoc, K., Zebrack, B. & Behar, L. (2003). A social work
guide to conducting research in psychosocial oncology. Philadelphia,
PA:Association of Oncology of Social Work.

Rubin, A., & Babbie, E. (1993). Research methods for social work (2nd ed.). Pacific
Grove, CA: Brooks/Cole.

Saunders, D.G. (1992). A typology of men who batter women: Three types derived
from cluster analysis. American Orthopsychiatry, 62, 264-275.

Sigelman, C.K., Berry, C.J., & Wiles, K.A. (1984). Violence in college students'
dating relationships. Journal of Applied Social Psychology, 5, 530-548.

Stets, J.E., & Henderson, D.A. (1991). Contextual factors surrounding conflict
resolution while dating: Results from a national study. Family Relations, 40,
29-36.

Stets, J.E., & Pirog-Good, M.A. (1987). Violence in dating relationships. Social
Psychological Quarterly, 50, 237-246.

Stith, S. M., Jester, S. B., & Bird, G. W. (1992). A typology of college students who
use violence in their dating relationships. Journal of College Student
Development, 33, 411-421.

Straus, M. A. (1979). Measuring intrafamily conflict and violence: The Conflict
Tactics (CT) Scales. Journal of Marriage and the Family, 41, 75-88.

116
Straus, M.A., & Gelles, R.J. (1990). Physical violence in American families: Risk
factors and adaptation to violence in 8,145 families. New Brunswick:
Transaction Publishers.

Straus, M.A., Gelles, R.J., & Steinmetz, S.K. (1980). Behind Closed Doors: Violence
in the American Family. Garden City, NY: Anchor.

Straus, M., & Hamby, S. (1997). Measuring physical and psychological
maltreatment of children with the Conflict Tactics Scales. In G. Kantor & J.
Jasinski (Eds.), Out of darkness: Contemporary perspectives on family
violence (pp. 119-135). Thousand Oaks, CA: Sage.

Stuart, G.L, Temple, J.R., & Moore, T.M. (2007). Improving batterer intervention
programs through theory-based research. JAMA, 298, 5, 560-563.

Thompson, T., & Massat, C.R. (2005). Experiences of violence, post-traumatic
stress, academic achievement, & behavior problems of urban african-
american children. Child & Adolescent Social Work Journal, 22, 5-6, 367-
393.

Tontodonato, P., & Crew, B.K. (1992). Dating violence, social learning theory, and
gender: A multivariate analysis. Violence and Victims, 7, 3-14.

Tweed, R., & Dutton, D. G. (1998). A comparison of impulsive and instrumental
subgroups of batterers. Violence and Victims, 13, 217-230.

Ward, J.H. (1963). Hierarchical grouping to optimize an objective function. Journal
of the American Statistical Association, 58, 236-244.

Wekerle, C. & Wolfe, D. (1998). The role of child maltreatment and attachment style
in adolescent relationship violence. Development & Psychopathology, 10(3),
571-586.

Widom, C.S. (2000). Motivation and mechanisms in the "cycle of violence." In D.J.
Hansen (Ed.), Nebraska Symposium on Motivation: Vol. 46. Motivation and
child maltreatment (pp. 1-37). Lincoln: University of Nebraska Press.

Wishart, D. (1969). An algorithm for hierarchical classifications. Biometrics, 25,
165-170.

Wishart, D. (1982). Supplement, CLUSTAN user manual (3
rd
ed.). Edinburgh,
Scotland: Program Library Unit, Edinburgh University.

117
Wolfe, D., Wekerle, C., Reitzel-Jaffe, D., & Lefebvre, L. (1998). Factors associated
with abusive relationships among maltreated and nonmaltreated youth.
Development & Psychopathology, 10(1), 61-85.

118
Appendix A
Please note: This questionnaire is completely anonymous. DO NOT put any
names on these pages. Also, you are free to refuse to answer any questions which
make you feel uncomfortable. Please follow instructions when answering this
questionnaire.

PART 1: Background Information
1. Age: _________ 2. Gender: Male_____ Female_____
3. Ethnicity: African American_____ White_____ Latino_____
Asian-Pacific_____ Native American_____ Other__________
4. Grade Level: 9th_____ 10th_____ 11th_____ 12th_____
5. Average Grade in School: A_____ B_____ C_____ D_____ F_____
6. Who lives at home with you: (Check all that apply)
_____ Mother _____ brothers/sisters -- How many: _____
_____ Father _____ Others living at home: _______________
_____ Stepfather _______________
_____ Stepmother _______________
7. Father’s education:
_____ no high school _____ some college
_____ some high school _____ college graduate
_____ high school graduate/GED _____ post college

119
8. Mother’s education:
_____ no high school _____ some college
_____ some high school _____ college graduate
_____ high school graduate/GED _____ post college
9. Father’s occupation: ____________________________________
10. Mother’s occupation: ____________________________________
11. Family total income:
_____ 0 - $15,000 _____ $45,001 - $60,000
_____ $15,001 - $30,000 _____ $60,001 and above
_____ $30,001 - $45,000 _____ unknown

PART 2: Information on Dating
1. At what age did you start dating: _____
2. How many boyfriends/girlfriends have you had since you began dating?
_____ only one _____ 5 to 10
_____ a few _____ more than 10
3. How frequently do you date?
_____ almost every night _____ once every 2 weeks
_____ 3 to 4 times a week _____ once a month
_____ 1 to 2 times a week _____ a few times per year
4. How many persons are you currently dating? __________

120
5. Some people think it is all right to hit or slap under certain conditions. After each
of the following statements indicate whether or not you think it is OK for couples to
slap of hit one another under the following situations. Circle a number from 1 to 4
that shows your opinion. For example:

1 = never all right, 2 = sometimes all right, 3 = usually all right, and 4 = always all
right.

Do you think it is all right for a girlfriend to
hit/slap boyfriend if:
Never
All
Right
Sometimes
All Right
Usually
All Right
Always
All Right
a- He threatens to break up with her 1 2 3 4
b- He is drunk/drugged and acting crazy 1 2 3 4
c- In an argument, he hits her first 1 2 3 4
d- She finds out he is cheating on her 1 2 3 4
e- He calls her nasty names 1 2 3 4
f- He flirts with another girl in front of her and
her friends at a party
1 2 3 4
g- He makes her look foolish in front of family
or friends
1 2 3 4
h- He refuses to let her go out for an evening
with her friends
1 2 3 4
i- Is there any other time it is OK?
Explain: 1 2 3 4
Do you think it is all right for a boyfriend to
hit/slap girlfriend if:
Never
All
Right
Sometimes
All Right
Usually
All Right
Always
All Right
a- She threatens to break up with him 1 2 3 4
b- She is drunk/drugged and acting crazy 1 2 3 4
c- In an argument, she hits him first 1 2 3 4
d- He finds out she is cheating on her 1 2 3 4
e- She calls him nasty names 1 2 3 4
f- She flirts with another boy in front of him and
his friends at a party
1 2 3 4
g- She makes him look foolish in front of family
or friends
1 2 3 4
h-She refuses to let him go out for an evening
with his friends
1 2 3 4
i- Is there any other time it is OK?
Explain: 1 2 3 4

121
6. Which statement best describes your beliefs about who should have control in
relationships: (Check one)
_____ The girl should have a lot of control over the boy.
_____ The girl should have a little control over the boy.
_____ Both the girl and boy should have equal control.
_____ The boy should have a lot of control over the girl.
_____ The boy should have a little control over the girl.
7. Using the following scale, indicate how true the following statements are for you:
1 = never 2 = seldom 3 = sometimes 4 = often 5 = very frequently
_____ It’s exciting to have a person demonstrate his/her physical power over you in
a way that is fun and not hurtful.
_____ It’s flattering to have a boyfriend/girlfriend who is jealous of you because it
shows how much he/she cares.
_____ In serious relationships almost all of one’s loyalty should be devoted to one’s
boyfriend/girlfriend rather than to one’s own friends.
_____ I keep my boyfriend/girlfriend in line.
_____ I am successful in getting my way with my boyfriend/girlfriend.
_____ I get my boyfriend/girlfriend to act in a way that I want him/her to act.
_____ My boyfriend/girlfriend is free to go where ever she/he wants to go even
though I may not approve.
_____ I accept my boyfriend/girlfriend for who he/she is rather than trying to
change him/her.

122
_____ When my boyfriend/girlfriend does not meet my expectation, I understand
and accept it.
8. Using the following scale, indicate whether you strongly agree, agree, disagree, or
strongly disagree with the following statements:
1 = Strongly agree 2 = Agree 3 = Disagree 4 = Strongly disagree
_____ On the whole, I am satisfied with myself.
_____ At times I think I am no good at all.
_____ I feel that I have a number of good qualities.
_____ I am able to do things as well as most other people
_____ I feel I do not have much to be proud of.
_____ I certainly feel useless at times.
_____ I feel that I’m a person of worth, at least on equal plane with others.
_____ I wish I could have more respect for myself.
_____ All in all, I am inclined to feel that I am a failure.
_____ I take a positive attitude toward myself.
9. Is there one person in your life you feel you can talk to about “almost
everything?” _____ yes _____ no
If yes, what is the relationship to that person? (Check all that apply)
_____ mother _____ sister
_____ father _____ brother
_____ boyfriend _____ girlfriend
_____ other: ___________________________________

123
PART 3: CONFLICTS IN DATING RELATIONSHIPS
Here is a list of things that might have happened when you have had a disagreement
with someone you have dated. Think about all present and past dating relationships
and how often your date did the acts to you and how often you did the acts to your
date. Please circle a number of times you or your date did these things:

Your Date Did To You
Never Once Twice 3-5 6-10 11-20 20+
times times times

Threw something at me 0 1 2 3 4 5 6
Pushed, grabbed, or shoved me 0 1 2 3 4 5 6
Slapped me 0 1 2 3 4 5 6
Forced me to do something sexual 0 1 2 3 4 5 6
Kicked, bit or hit me with a fist 0 1 2 3 4 5 6
Hit or tried to hit me with something 0 1 2 3 4 5 6
Beat me up 0 1 2 3 4 5 6
Threatened me with a knife or gun 0 1 2 3 4 5 6
Used a knife or gun on me 0 1 2 3 4 5 6

You Did To Your Date
Never Once Twice 3-5 6-10 11-20 20+
times times times

Threw something at him/her 0 1 2 3 4 5 6
Pushed, grabbed, or shoved him/her 0 1 2 3 4 5 6
Slapped him/her 0 1 2 3 4 5 6
Forced him/her to do something sexual 0 1 2 3 4 5 6
Kicked, bit or hit with a fist 0 1 2 3 4 5 6
Hit or tried to hit with something 0 1 2 3 4 5 6

124
Beat up 0 1 2 3 4 5 6
Threatened with a knife or gun 0 1 2 3 4 5 6
Used a knife or gun 0 1 2 3 4 5 6

If you answered 0 on all above items, skip to PART 5. If not, go to next
question, PART 4.

PART 4: Answer these questions if you circled 1 or more items on the last
section.

1. In how many relationships did any of the above aggressive acts occur? _____

If the acts of physical aggression occurred in more than one relationship, answer the
next set of questions with regard to the last relationship in which they occurred.

2. Are you presently dating that same person? _____ yes _____ no
3. How long have you dated that person? _____ weeks/months/years
4. How serious is or was the relationship? (Check one)
_____ casual date
_____ someone you have dated often but are not emotionally attached to
_____ someone you are emotionally attached to but not in love
_____ someone you are in love with
_____ someone you are in love with and would like to marry, not yet discussed
marriage
_____ someone you are in love with, have discussed marriage, but made no plans
_____ someone you are in love with and have made plans to marry

125
5. Please CIRCLE the number on the scale that best describes or described your
relationship with your boyfriend/girlfriend. Circle only one number for each item.
How well does/did your boyfriend/girlfriend meet your needs?
1= Not at all 2 3 4 5 = extremely well
In general, how satisfied are/were you with your relationship?
1= Unsatisfied 2 3 4 5 = extremely satisfied
How good is/was your relationship compared to most?
1= Poor 2 3 4 5 = excellent
How often do/did you wish you hadn’t gotten into this relationship?
1= Never 2 3 4 5 = Very often
To what extent has your relationship met your original expectations?
1= Hardly at all 2 3 4 5 = Completely
How much do/did you love your boyfriend/girlfriend?
1= Not much 2 3 4 5 = Very much
How many problems are/were there in your relationship?
1= Very few 2 3 4 5 = Very many

6. At what stage did the first physical aggressive act occur? (Check one)
_____ first date _____ steady or serious dating
_____ casual dating _____ engaged

126
7. Which of the following statements best describes who threatened to hit first,
started to push or shove or throw the first slap or punch? (Check one)
_____ My boyfriend/girlfriend always started the violence.
_____ My boyfriend/girlfriend usually started the violence.
_____ My boyfriend/girlfriend and I were equally responsible for starting the
violence.
_____ I usually started the violence.
_____ I always started the violence.

8. Using the following scale, indicate how true the following statements are for you:
1 = never 2 = seldom 3 = sometimes 4 = often 5 = very frequently

_____ How often did you and your boyfriend/girlfriend argue with each other?
_____ How often did you try to change things about your boy/girlfriend that
bothered you?
_____ How often did you feel angry or resentful toward your boyfriend/girlfriend?
_____ When you and your boyfriend/girlfriend argued, how serious were the
arguments?
_____ How often did you communicate negative feelings toward your
boyfriend/girlfriend, for example, anger, dissatisfaction, frustration?

127
Think about the last time you experienced physical aggression in your
relationship

9. What were the main reasons for your partner’s use of physical aggression in the
relationship? (Check the main 2 reasons. Write 1 for the first and 2 for the second)
_____ to show anger _____ in self defense
_____ was unable to express self verbally _____ to punish me for wrong behavior
_____ to feel more powerful _____ to prove love
_____ to get control over me _____ to get attention
_____ to get back at me _____ jealousy
_____ was mad at someone/thing else _____ drug alcohol
_____ other_________________________________________________________

10. What were the main reasons for your use of physical aggression in the
relationship? (Check the main 2 reasons. Write 1 for the first and 2 for the second)
_____ to show anger _____ in self defense
_____ was unable to express self verbally _____ to punish person for wrong
behavior
_____ to feel more powerful _____ to prove love
_____ to get control over other person _____ to get attention
_____ to get back at him/her _____ jealousy
_____ was mad at someone/thing else _____ drug alcohol
_____ other_________________________________________________________

128
11. How did you react to your partner’s physical aggression toward you? (Check all
that apply)
_____ fear _____ sad/depressed _____ hopelessness
_____ anger _____ ignored it _____ afraid of not being loved
_____ emotionally hurt _____ felt guilty _____ felt partner loved me
_____ thought it was funny _____ no effect at all
_____ shame / humiliation _____ wanted to get revenge
_____ wanted to get away _____ felt I deserved it
_____ other reactions: _________________________________________________

12. How do you think your partner reacted to the physical aggression you did to
him/her? (Check all that apply)
_____ fear _____ sad/depressed _____ hopelessness
_____ anger _____ ignored it _____ afraid of not being loved
_____ emotionally hurt _____ felt guilty _____ felt partner loved me
_____ though it was funny _____ no effect at all
_____ shame / humiliation _____ wanted to get revenge
_____ wanted to get away _____ felt I deserved it
_____ other reactions: __________________________________________________

129
13. With whom did you talk about the physical aggression?
_____ no one _____ teacher
_____ friend _____ other: ____________________
_____ family

14. What was the effect of the physical aggression on the relationship? (Check one)
_____ improved _____ stayed the same
_____ got worse _____ relationship ended

After completing this section, skip PART 5 and go to PART 6.

PART 5: Skip this section if you completed PART 4.
If you did not complete PART 4, answer the following questions with regard to
your present or most recent relationship.

1. Is this a present or past dating relationship? _____ present _____ past
2. How long have you dated your boyfriend/girlfriend? _____ weeks/months/years
3. How serious is the relationship? (Check one)
_____ casual date
_____ someone you have dated often but are not emotionally attached to
_____ someone you are emotionally attached to but not in love
_____ someone you are in love with

130
_____ someone you are in love with and would like to marry, not yet discussed
marriage
_____ someone you are in love with, have discussed marriage, but made no plans
_____ someone you are in love with and have made plans to marry

4. Please CIRCLE the number on the scale that best describes or described your
relationship with your boyfriend/girlfriend. Circle only one number for each item.
a- How well does/did your boyfriend/girlfriend meet your needs?
1= Not at all 2 3 4 5 = extremely well
In general, how satisfied are/were you with your relationship?
1= Unsatisfied 2 3 4 5 = extremely satisfied
How good is/was your relationship compared to most?
1= Poor 2 3 4 5 = excellent
How often do/did you wish you hadn’t gotten into this relationship?
1= Never 2 3 4 5 = Very often
To what extent has your relationship met your original expectations?
1= Hardly at all 2 3 4 5 = Completely
How much do/did you love your boyfriend/girlfriend?
1= Not much 2 3 4 5 = Very much
How many problems are/were there in your relationship?
1= Very few 2 3 4 5 = Very many

131
5. Using the following scale, indicate how true the following statements are for you:
1 = never 2 = seldom 3 = sometimes 4 = often 5 = very
frequently
_____ How often did you and your boyfriend/girlfriend argue with each other?
_____ How often did you try to change things about your boy/girlfriend that
bothered you?
_____ How often did you feel angry or resentful toward your boyfriend/girlfriend?
_____ When you and your boyfriend/girlfriend argued, how serious were the
arguments?
_____ How often did you communicate negative feelings toward your
boyfriend/girlfriend, for example, anger, dissatisfaction, frustration?

132
PART 6: CONFLICTS IN PARENTS RELATIONSHIP
1. Here is a list of things PARENTS might have done when they have had a conflict
or disagreement with each other. PARENTS refer to any natural, adoptive, or
stepparents you have lived with. Think about the year your parents seemed to
argue the most. Please circle a number for each of the items listed below to show
the total number of times your mother/stepmother or father/stepfather did them in
that year.

Father Did To Mother
Never Once Twice 3-5 6-10 11-20 20+
times times times

Threw something at the other 0 1 2 3 4 5 6
Pushed, grabbed, or shoved 0 1 2 3 4 5 6
Slapped the other 0 1 2 3 4 5 6
Kicked, bit or hit with a fist 0 1 2 3 4 5 6
Hit or tried to hit with something 0 1 2 3 4 5 6
Beat up 0 1 2 3 4 5 6
Threatened with a knife or gun 0 1 2 3 4 5 6
Used a knife or gun 0 1 2 3 4 5 6

Mother Did To Father
Never Once Twice 3-5 6-10 11-20 20+
times times times

Threw something at the other 0 1 2 3 4 5 6
Pushed, grabbed, or shoved 0 1 2 3 4 5 6
Slapped the other 0 1 2 3 4 5 6
Kicked, bit or hit with a fist 0 1 2 3 4 5 6

133
Hit or tried to hit with something 0 1 2 3 4 5 6
Beat up 0 1 2 3 4 5 6
Threatened with a knife or gun 0 1 2 3 4 5 6
Used a knife or gun 0 1 2 3 4 5 6

CONFLICTS BETWEEN YOU AND YOUR PARENTS
2. Here is a list of things YOUR PARENTS might have done when they have had a
conflict or disagreement with YOU when you were growing up. PARENTS refer to
any natural, adoptive, or stepparents you have lived with. Think about the year
you and your parents seemed fought the most. Please circle a number for each of
the items listed below to show the total number of times your mother/stepmother or
father/stepfather did them.

Father Did To Me
Never Once Twice 3-5 6-10 11-20 20+
times times times

Threw something at the me 0 1 2 3 4 5 6
Pushed, grabbed, or shoved me 0 1 2 3 4 5 6
Slapped/spanked me 0 1 2 3 4 5 6
Kicked, bit or hit me with a fist 0 1 2 3 4 5 6
Hit or tried to hit me with something 0 1 2 3 4 5 6
Ex. belt
Beat me up 0 1 2 3 4 5 6
Threatened me with a knife or gun 0 1 2 3 4 5 6
Used a knife or gun 0 1 2 3 4 5 6
Other: ______________________ 0 1 2 3 4 5 6

134
Mother Did To Me
Never Once Twice 3-5 6-10 11-20 20+
times times times

Threw something at me 0 1 2 3 4 5 6
Pushed, grabbed, or shoved me 0 1 2 3 4 5 6
Slapped/spanked me 0 1 2 3 4 5 6
Kicked, bit or hit me with a fist 0 1 2 3 4 5 6
Hit or tried to hit me with something 0 1 2 3 4 5 6
Beat me up 0 1 2 3 4 5 6
Threatened me with a knife or gun 0 1 2 3 4 5 6
Used a knife or gun 0 1 2 3 4 5 6
Other: ______________________ 0 1 2 3 4 5 6

3. Have you ever seen acts of aggression/violence in your neighborhood?
_____ yes _____ no
If yes, indicate by checking the type and frequency of the violence you have seen
during the past year. (Check only those that apply)
Daily Weekly Monthly 2x/year 1x/year
pushing, grabbing, shoving ___ ___ ___ ___ ___
kicking, hitting with fist ___ ___ ___ ___ ___
hitting with object ___ ___ ___ ___ ___
beating up ___ ___ ___ ___ ___
threatening with knife or gun ___ ___ ___ ___ ___
stabbing ___ ___ ___ ___ ___

135
shooting ___ ___ ___ ___ ___
burglary ___ ___ ___ ___ ___
vandalism (destroying property) ___ ___ ___ ___ ___
rape ___ ___ ___ ___ ___
Other: ______________________ ___ ___ ___ ___ ___

4. Have you ever seen acts of aggression/violence in your school?
_____ yes _____ no
If yes, indicate by checking the type and frequency of the violence you have seen
during the past year. (Check only those that apply)
Daily Weekly Monthly 2x/year 1x/year
pushing, grabbing, shoving ___ ___ ___ ___ ___
kicking, hitting with fist ___ ___ ___ ___ ___
hitting with object ___ ___ ___ ___ ___
beating up ___ ___ ___ ___ ___
threatening with knife or gun ___ ___ ___ ___ ___
stabbing ___ ___ ___ ___ ___
shooting ___ ___ ___ ___ ___
burglary ___ ___ ___ ___ ___
vandalism (destroying property) ___ ___ ___ ___ ___
rape ___ ___ ___ ___ ___
Other: ______________________ ___ ___ ___ ___ ___

136
5. How often, if ever, did you get into a physical fight with some other person during
high school?
_____ never _____ sometimes
_____ seldom _____ often

6. How often, if ever, did you get into a physical fight with some other person during
junior high school?
_____ never _____ sometimes
_____ seldom _____ often

7. How often, if ever, did you get into a physical fight with some other person during
elementary school?
_____ never _____ sometimes
_____ seldom _____ often

8. How often, if ever, do you drink alcohol?
_____ almost daily _____ about once every 6 months
_____ almost weekly _____ no more than once a year
_____ about once a month _____ never

137
9. How often, if ever, do you smoke marijuana?
_____ almost daily _____ about once every 6 months
_____ almost weekly _____ no more than once a year
_____ about once a month _____ never

10. How often, if ever, do you do other drugs?
_____ almost daily _____ about once every 6 months
_____ almost weekly _____ no more than once a year
_____ about once a month _____ never

Below is a list that describes kids. For each item that describes you now or within
the past 6 months, please circle the 2 if the item is very true or often true of you.
Circle the 1 if the item is somewhat or sometimes true of you. If the item is not
true of you, circle the 0.

2 = very true or often true 1 = somewhat or sometimes true 0 = not true
I act too young for my age 0 1 2
I have an allergy 0 1 2
Describe____________________________
I argue a lot 0 1 2
I have asthma 0 1 2
I act like the opposite sex 0 1 2
I like animals 0 1 2
I brag 0 1 2

138
I have trouble concentrating or paying attention 0 1 2
I can’t get my mind off certain thoughts 0 1 2
Describe_____________________________
I have trouble sitting still 0 1 2
I’m too dependent on adults 0 1 2
I feel lonely 0 1 2
I feel confused or in a fog 0 1 2
I cry a lot 0 1 2
I am pretty honest 0 1 2
I am mean to others 0 1 2
I daydream a lot 0 1 2
I deliberately try to hurt or kill myself 0 1 2
I try to get a lot of attention 0 1 2
I destroy my own things 0 1 2
I destroy things belonging to others 0 1 2
I disobey my parents 0 1 2
I disobey at school 0 1 2
I don’t eat as well as I should 0 1 2
I don’t get along with other kids 0 1 2
I don’t feel guilty when doing something I shouldn’t 0 1 2
I am jealous of others 0 1 2
I am willing to help others when they need help 0 1 2

139
I am afraid of certain animals, situations, and places 0 1 2
Describe_____________________________
I am afraid of going to school 0 1 2
I am afraid that I might think of do something bad 0 1 2
I feel that I have to be perfect 0 1 2
I feel that no one loves me 0 1 2
I feel that others are out to get me 0 1 2
I feel worthless or inferior 0 1 2
I accidentally get hurt a lot 0 1 2
I get in may fights 0 1 2
I get teased a lot 0 1 2
I hang around with kids that get into trouble 0 1 2
I hear sounds or voices that other people don’t 0 1 2
Describe_____________________________
I act without stopping to think 0 1 2
I would rather be alone than with others 0 1 2
I lie or cheat 0 1 2
I bite my fingernails 0 1 2
I am nervous or tense 0 1 2
Parts of my body twitch or make nervous movements 0 1 2
Describe_____________________________
I have nightmares 0 1 2

140
I am not liked by other kids 0 1 2
I can do certain things better than most kids 0 1 2
I am too fearful or anxious 0 1 2
I feel dizzy 0 1 2
I feel too guilty 0 1 2
I eat too much 0 1 2
I feel overtired 0 1 2
I am overweight
Physical problems without known medical cause 0 1 2
Aches or pains (not headaches) 0 1 2
Headaches 0 1 2
Nausea, feel sick 0 1 2
Problems with eyes 0 1 2
Describe_____________________________
Rashes or other skin problems 0 1 2
Stomachaches or cramps 0 1 2
Vomiting, throwing up 0 1 2
I physically attack people 0 1 2
I pick my skin or other parts of my body 0 1 2
Describe_____________________________
I can be pretty friendly 0 1 2
I like to try new things 0 1 2

141
My school work is poor 0 1 2
I am poorly coordinated or clumsy 0 1 2
I would rather be with older kids than with kids my own age 0 1 2
I would rather be with younger kids than with kids my own age 0 1 2
I refuse to talk 0 1 2
I repeat certain actions over and over 0 1 2
Describe_____________________________
I run away from home 0 1 2
I scream a lot 0 1 2
I am secretive or keep things to myself 0 1 2
I see things that other people think aren’t there 0 1 2
Describe_____________________________
I am self-conscious or easily embarrassed 0 1 2
I set fires 0 1 2
I can work well with my hands 0 1 2
I show off or clown 0 1 2
I am shy 0 1 2
I sleep less than most kids 0 1 2
I sleep more than most kids during the day and/or night 0 1 2
Describe_____________________________
I have a good imagination 0 1 2
I have a speech problem 0 1 2

142
Describe_____________________________
I stand up for my rights 0 1 2
I steal at home 0 1 2
I steal from other places other than home 0 1 2
I store things up I don’t need 0 1 2
Describe_____________________________
I do things other people think are strange 0 1 2
Describe_____________________________
I have dreams that other people would think are strange 0 1 2
Describe_____________________________
I am stubborn 0 1 2
My moods or feelings change suddenly 0 1 2
I enjoy being with other people 0 1 2
I am suspicious 0 1 2
I swear or use dirty language 0 1 2
I think about killing myself 0 1 2
I like to make others laugh 0 1 2
I talk too much 0 1 2
I tease others a lot 0 1 2
I have a hot temper 0 1 2
I think about sex 0 1 2
I threaten to hurt people 0 1 2

143
I like to help others 0 1 2
I am concerned about being neat or clean 0 1 2
I have trouble sleeping 0 1 2
Describe_____________________________
I cut classes or skip school 0 1 2
I don’t have much energy 0 1 2
I am unhappy, sad, or depressed 0 1 2
I am louder than other kids 0 1 2
I use alcohol or drugs for nonmedical purposes 0 1 2
Describe_____________________________
I try to be fair to others 0 1 2
I enjoy a good joke 0 1 2
I like to take life easy 0 1 2
I try to help other people when I can 0 1 2
I wish I were of the opposite sex 0 1 2
I keep from getting involved with others 0 1 2
I worry a lot 0 1 2
Please write down anything else that describes your feelings, behavior, or
interests:

Abstract (if available)

Abstract The study s purpose was twofold: (i) identify subtypes for adolescent male and female perpetrators of dating violence, and (ii) examine distal and proximal risk factors of adolescent dating violence for differing male and female typologies. A cluster analysis was used to develop perpetrator typologies: 3-clusters emerged for males -- generally violent

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

School social dynamics as mediators of students personal traits and family factors on the perpetration of school violence in Taiwan

PDF

Examining the longitudinal relationships between community violence exposure and aggressive behavior among a sample of maltreated and non-maltreated adolescents

PDF

Complicated relationships: the prevalence and correlates of sexual assault, dating violence victimization, and minority stress among sexual minority adolescents throughout the United States

PDF

Healthy relationship education using applied academics: a solution to decreasing teen dating violence among high school students

PDF

Untangling the developmental relations between depression and externalizing behavior among maltreated adolescents

PDF

An examination of sexual harassment, gender discrimination, stalking, and sexual assault among female and male veterans and associations with PTSD and depression

Asset Metadata

Creator Thomson, Christopher Michael (author)

Core Title Typologies of male and female adolescent perpetrators of dating violence: three subtypes for males and four subtypes for females, and differences among them

School School of Social Work

Degree Doctor of Philosophy

Degree Program Social Work

Publication Date 02/12/2010

Defense Date 12/11/2007

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag dating violence,domestic volence,gender,generally violent,OAI-PMH Harvest,perpetrators,risk factors,typologies

Language English

Advisor Zebrack, Brad J. (committee chair), Moore, Robert S. (committee member), O'Keefe, Maura (committee member)

Creator Email cmthoms@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m1011

Unique identifier UC1318821

Identifier etd-Thomson-20080212 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-41675 (legacy record id),usctheses-m1011 (legacy record id)

Legacy Identifier etd-Thomson-20080212.pdf

Dmrecord 41675

Document Type Dissertation

Rights Thomson, Christopher Michael

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu