[acb-hsp] Readiness to Change
J.Rayl
thedogmom63 at frontier.com
Mon Jul 25 22:34:45 EDT 2011
I hope all of this comes through.
Readiness to Change as a Predictor of Outcome in Batterer Treatment
Katreena L. Scott and David A. Wolfe
University of Western Ontario
The current study examined stage of change as a predictor of outcome in batterer treatment. Men (N
119) were classified into the transtheoretical model's stages of change and assessed 3 times over
treatment. Hierarchical linear modeling revealed significant variation in men's progress, predictable from
their stage of change. As hypothesized, men in the precontemplation stage showed little positive change
in empathy, communication, or abusive behavior, whereas men in the contemplation and action stages
showed positive growth in all of these domains. These effects occurred in the initial 10 weeks of
treatment, after which men progressed at a more homogeneous rate. Interpretation is complicated by
pretreatment differences that draw into question stage-related patterns in final outcome. Implications for
general models of abuse cessation and for stage-specific trajectories are discussed.
Treatment programs for men who are physically, verbally, and
emotionally abusive toward their intimate partners began in the
mid-1970s and spread rapidly across North America and Europe
(Roberts, 1984). There are now hundreds of programs offering
services specifically for this client population (National Clearinghouse
on Family Violence, 2002). Most programs operate according
to a set of regulatory standards (Austin & Dankwort, 1999),
and many are integrated into the wider criminal justice response to
domestic violence (Murphy, Musser, & Maton, 1998).
Despite this growth in popularity, the efficacy of batterer treatment
remains uncertain. Rates of dropout from batterer programs
are high, typically varying between 30% and 60% (Daly & Pelowski,
2000). Among men who complete treatment, approximately
one third engage in physically violent behavior following
program completion (Scott, 2002), a rate of success that is inferior
to treatment for general mental health problems, such as depression,
anxiety, and marital distress (e.g., Johnson, Hunsley, Greenberg,
& Schindler, 1999). Moreover, controversy continues over
whether men randomly assigned to treatment reassault their partners
at a lower rate than nonintervention controls (compare Dunford,
2000, with Gondolf, 2002).
In this context of high dropout rates and unconvincing evidence
of treatment efficacy, an important question is the identification of
men who, even under ideal conditions, are most and least likely to
benefit from treatment. Although not relevant to judging the overall
efficacy of batterer treatment, identifying clients for whom
traditional treatment can be successful permits the best use of
resources for encouraging treatment attendance and promoting
change. As well, such identification prompts consideration of
possible alternative interventions for men who are not successful
in traditional programs.
Consistent with this reasoning, there are a growing number of
studies on predictors of success in batterer intervention. Prior
studies have examined demographic variables (e.g., men's age,
education, and ethnicity), pretreatment relationship characteristics
(e.g., severity of past abuse and relationship satisfaction), and
concurrent presenting problems (e.g., men's scores on depression
inventories) without notable success (e.g., Faulkner, Stoltenberg,
Cogen, Nolder, & Shooter, 1992; Gondolf, 1997, 2002; Hamberger
& Hastings, 1988). Results of studies on the predictive value of
psychopathology and of personality typologies have been more
promising but still yield mixed results. For example, antisocial/
psychopathic traits were predictive of posttreatment recidivism in
a study by Dutton, Bodnarchuk, Kropp, Hart, and Ogloff (1997),
but they were unrelated to recidivism in a study by Gondolf
(2002). Results of studies using a trichotomous division of men
into family-only, dysphoric/borderline, and generally violent/antisocial
batterers are also mixed (Gondolf, 2002, vs. Saunders,
1996). Perhaps the most consistent association with treatment
success is men's alcohol use, with more frequent and severe
posttreatment violence associated with men's drinking (Fals-
Stewart, Kashdan, O'Farrell, & Birchler, 2002; Hamberger &
Hastings, 1990), a relationship that has implications for adding to,
but not necessarily for modifying, current treatment targets of
batterer treatment programs.
The current research represents the culmination of a series of
studies (Scott & Wolfe, 2000, 2002, in press) designed to examine
a strategy for predicting change among batterers based on the
transtheoretical model (TTM; Prochaska, DiClemente, &
Norcross, 1992). The TTM was initially developed to aid in
understanding and predicting change in health behaviors, such as
cigarette smoking, regular exercise, and condom use (Prochaska et
al., 1994). At the backbone of this model is a division of individ-
Katreena L. Scott and David A. Wolfe, Department of Psychology,
University of Western Ontario, Ontario, Canada.
David A. Wolfe is now at the Centre for Addiction and Mental Health,
University of Toronto, Ontario, Canada.
This research was supported by Doctoral Fellowship 752-97-1276 from
the Social Sciences and Humanities Research Council of Canada. We
acknowledge Changing Ways, London, and Tim Kelly for facilitating this
research; Kathy Pitulko and M. T. McNabb for their assistance in data
collection and aspects of analyses; and Liora Swartzman, William Fisher,
and members of the Family Violence Research Lab for their helpful
comments on a draft of this article.
Correspondence concerning this article should be addressed to Katreena
L. Scott, who is now at the Department of Human Development and
Applied Psychology, Ontario Institute for Studies in Education, University
of Toronto, 252 Bloor Street West, Toronto, Ontario M5S 1V6, Canada.
E-mail: kscott at oise.utoronto.ca
Journal of Consulting and Clinical Psychology Copyright 2003 by the American Psychological Association, Inc.
2003, Vol. 71, No. 5, 879-889 0022-006X/03/$12.00 DOI: 10.1037/0022-006X.71.5.879
879
uals into four subgroups or stages, homogeneous in terms of their
change attitudes and behaviors. The first group, labeled precontemplation,
includes individuals who deny the need to change and
who are not engaging in any change behaviors. The second group,
contemplation, includes individuals who intend to change but are
not yet changing, and the third group, action, includes individuals
who are engaged in serious change attempts. Individuals in the
final group, maintenance, are primarily concerned with retaining
changes they have made. Success in making intentional changes is
hypothesized to result from sequential, although not necessarily
linear, movement through each of these stages. The fundamental
propositions of the TTM are quite similar to stage-based conceptualizations
presented in the batterer intervention literature during
the mid 1980s (Adams, 1985; Gondolf, 1987).
Although the concept of change stages is quite popular in
clinical settings, a number of researchers have questioned whether
the TTM stages are qualitatively distinct or merely points along a
continuum (Sutton, 2001; Weinstein, Rothman, & Sutton, 1998).
Nonetheless, from an applied standpoint, it is recognized that
different interventions and prognoses may be associated with
individuals who differ qualitatively or quantitatively on a particular
dimension (e.g., different interventions are associated with
depression of greater or lesser severity). A more important question
may be the identification of the dimension(s) being assessed
by continuous and algorithmic stage-of-change measures. An individual's
stage of change, or "readiness to change," likely reflects
a number of underlying constructs, including change motivation,
efficacy, denial, and openness to seeking help (L. Jones, personal
communication, February 22, 2002).
Acknowledgement that the stages of change may be measuring
a number of familiar constructs may be a strength that ties this
model to other work on the relevance of attitudes toward change to
treatment outcome. Common factor models of psychotherapy identify
client factors, such as change expectancy and motivation, as
key predictors of treatment success (e.g., Kirsch, 1999). Psychodynamic
theorists also emphasize the role of denial in batterers,
linking it to projective identification, one of a number of possible
unconscious defenses against negative affect, such as shame or
trauma (Scalia, 1994).
The potential usefulness of the TTM in populations of batterers
was discussed in the mid-1990s by Dutton (1995) and Daniels and
Murphy (1997). Subsequent studies have supported the construct
and concurrent validity of the TTM in this population (e.g.,
Levesque, Gelles, & Velicer, 2000; Scott & Wolfe, 2002). For
example, cross-sectional studies have found that men inductively
classified in the precontemplation stage report lower levels of
abusive behavior than men in the contemplation and action stages,
despite similarly maladaptive scores on objective and external
problem indicators (Murphy, Begun, & Strodthoff, 1999; Scott &
Wolfe, 2000). Ongoing work by a number of researchers (e.g.,
Begun et al., 2003; Taft, Murphy, King, Musser, & DeDeyn, 2003)
continues the investigation of stage of change in application to
batterers.
Although promising, the most critical aspect of TTM-the value
of stage-of-change classifications for predicting abuse cessation-
has not been established. Treatment outcome studies of other
problematic behaviors, particularly smoking, have established that
individuals in later stages of change are more likely to change
problematic behaviors than those in earlier stages (e.g., Ockene et
al., 1992; Pallonen et al., 1994). Applied to a sample of batterers,
this pattern of results suggests that few abusive men who enter
treatment in the precontemplation stage will reduce their abusive
behavior. In contrast, men in the contemplation and action stages
should show progressively greater reductions in their abuse over
time.
In summary, although there are some outstanding theoretical
issues with the definition of constructs assessed by the TTM, there
is ample theoretical and empirical justification for the hypothesis
that individuals who deny having problems and who do not desire
change-in other words, those in a lower stage of change-will
show less behavior change over treatment than individuals with
higher levels of change readiness. The current study examines this
proposition with a relatively large sample of batterers over three
time points. Instead of focusing solely on abusive behavior as the
outcome of interest, this investigation was broadened to include a
range of variables that are theoretically expected to promote and
maintain changes in men's abusive behavior. These variables
included gains in empathy, communication skills, and perceptions
of personal responsibility for abusive behavior. We hypothesized
that men, on average, would make moderate reductions in abusive
behavior and moderate gains in all other variables over treatment
but that this average effect would obscure substantial interindividual
variations in progress. Men's stage of change at treatment
initiation was expected to be a significant predictor of this variation
after controlling for sociodemographic (e.g., age), relationship
characteristics (i.e., contact with partner, stability of relationship),
and concurrent risk factors (i.e., alcoholism, legal involvement).
Specifically, men in the precontemplation stage were expected to
have the lowest rate of change, men in the action stage the greatest
rate of change, and men in the contemplation stage a moderate rate
of change across all outcome domains.
Method
Participants
Participants were 119 men who completed treatment at Changing Ways,
London, a treatment agency located in a midsized city in southwestern
Ontario devoted to preventing and ending men's abuse of women. Men
were drawn from an initial sample of 345 men who attended an information
session at the agency between March 1998 and May 1999. Of these men,
9% (n 32) declined participation in the current study. An additional 86
men (25%) came to the orientation session but failed to attend more than
one intervention session, and 108 men (35%) dropped out at some time
later in treatment. This dropout rate, although high, is consistent with the
literature on attrition in batterer intervention programs (Daly & Pelowski,
2000). Because this study excluded dropouts, results are generalizable only
to men who complete treatment programs. A systematic comparison of the
194 men who failed to complete treatment and the 119 men who completed
treatment is detailed in Scott and Wolfe (in press). Briefly, there were few
differences in the characteristics of men who completed and dropped out of
treatment. Dropouts were more likely to be attending treatment voluntarily,
to have been arrested in the past, and to be younger. After controlling for
these differences, counselor-rated, but not self-reported, stage of change
added to the prediction of dropout; however, the best combination of
variables predicted less than 10% of the variance in men's treatment
completion status. The low predictive value of individual characteristics in
this study (and others) emphasizes the need for continued research in this
area and for creativity in considering contextual and interactional variables
that may relate to men's attrition (Daly & Pelowski, 2000).
880 SCOTT AND WOLFE
The remaining 119 men were predominantly Caucasian1 (90%) and
ranged in age from 23 to 70 years, with an average age of 37.5 years
(SD 9.34). Most men completed questionnaires at treatment initiation
and termination (108 and 106, respectively), with slightly lower completion
rates at treatment midpoint (n 92). Almost two thirds of the men in
this sample were attending treatment to fulfill the requirements of a
probation order (64%), with others attending most often in response to
pressure from their partners, friends, employers, or child protective services.
Most (80%) were living with their partners at treatment initiation,
with the remaining 20% divided between those who had daily contact with
their partners (10.5%) and those with less frequent partner contact (9.2%).
Men's relationships were generally stable over treatment (76%), although
a few resumed (14%) or left (10%) a relationship over this time. Men were
characterized by numerous indicators of lifestyle instability (Gondolf,
2002). Over one third (40%) had not completed high school, approximately
one quarter (27%) were not employed, and almost half had a history of
prior arrest (45%).
In addition to participants' self-report, we made an attempt to gather
collateral data from men's partners. Because men whose partners have left
them often pursue, threaten, and assault their former wives and girlfriends,
both estranged and current partners were considered eligible for research
participation. Unfortunately, research information was available for
only 35 women (16 at the initiation, 20 at the midpoint, and 17 at the
termination of their partner's treatment). Of the 84 women not contacted,
76% were ineligible because they were not contacted by or they refused
involvement with the treatment agency's Battered Woman's Advocate;2
6% were eligible but declined participation in this study; and 18% consented
but could not be subsequently located by the researchers. Because
of the relatively small number of women remaining, a potential bias toward
women experiencing less severe forms of abuse was a significant concern
(Petrik, Gildersleeve-High, McEllistrem, & Subotnik, 1994). To examine
this potential bias, we compared partners of women who were and were not
contacted on male-reported abusive behavior, level of responsibility for
abuse, referral source, arrest history, relationship status, and stage of
change. No significant differences and no concerning trends were found in
these comparisons in multivariate analysis of variance, F(10, 56) .12, p
.05, tentatively supporting the null hypothesis that partner contact was
not limited to those women least victimized by their partners. Nevertheless,
caution is warranted in the interpretation of women's reports of their
partners' behavior.
Measures
Sociodemographic and relationship characteristics and concurrent risk
variables. Men provided information on their age, ethnicity, education,
employment status, history of arrest, and referral source (court ordered or
voluntary) on simple self-report questions during their first assessment. In
addition, history of victimization was assessed with two questions adapted
from the National Family Violence Survey (NFVS; Straus & Gelles, 1990).
The first question asked men to report on the frequency with which they
experienced physical punishment during adolescence, and the second asked
whether they had witnessed domestic violence. For men who did not
self-identify as alcoholic, problems with alcohol were assessed with the
Short Michigan Alcoholism Screening Test (Selzer, Vinokur, & van Rooijen,
1975). Men's self-identification and responses on the Short Michigan
Alcoholism Screening Test were used to classify them into three groups:
nonalcoholic, possibly alcoholic, and alcoholic. Finally, men responded to
questions about their relationship status (i.e., living together, not living
together but daily contact, and not living together and less than daily
contact) at each assessment point.
Stage of change. Stage-of-change classifications were made based on
the University of Rhode Island Change Assessment Scale (McConnaughy,
DiClemente, Prochaska, & Velicer, 1989). This scale asks men to rate their
agreement on a 5-point Likert scale with eight statements reflecting each of
the precontemplation (e.g., "I am not the problem one, it doesn't make
sense for me to be here"), contemplation (e.g., "I have a problem and I
really think I should work on it"), action (e.g., "I am finally doing some
work on my problem"), and maintenance stages (e.g., "It worries me that
I might slip back on a problem I already have, so I am here to seek help").
Explicit mention was not made of the nature of the men's problem,
although all were aware that they were attending a program that exclusively
services men who are abusive in their intimate relationships. This
stage-of-change assessment strategy was chosen over measures specific to
intimate partner violence (e.g., Levesque et al., 2000) because many men
reject the label abuse for their behaviors, even if they are relatively
motivated to attend treatment. Men's responses were summed for each
scale, resulting in four subscales with adequate internal consistency (
.81, .87, .85, and .82, for precontemplation, contemplation, action, and
maintenance, respectively). Cluster analysis of men's responses to the
University of Rhode Island Change Assessment Scale suggested three
distinct groups corresponding to the precontemplation (n 43, 41%),
contemplation (n 36, 35%), and action (n 25, 24%) stages,3 as detailed
in Scott and Wolfe (2002).
Self-reported abuse perpetration. Men were asked to report on the
incidence (occurrence over the previous 2 months) of 31 physically and
emotionally abusive behaviors. To reduce potential bias in time frame of
reporting, we instructed men to consider an event that occurred 2 months
ago (e.g., a birthday) and base their reports on events that had occurred
since that time. The abuse inventory consisted of the 8 physical abuse items
from the Conflict Tactics Scale (Straus, 1979) and 15 items from an early
version of the Psychological Maltreatment of Women Inventory, Short
Version (Tolman, 1999; 7 assessing dominance and isolation and 8 assessing
emotional/verbal abuse). In addition, 4 items were added to assess
men's threatening (e.g., "threatened to leave the relationship") and sexually
abusive behaviors (e.g., "pressured your partner to have sex after a fight").
The latter items were compiled from several scales that attempted to assess
the full range of abusive behaviors. Although men initially provided
information on both the incidence and frequency of their abusive behaviors,
responses were later collapsed so that each item was scored in a yes/no
manner and summed across items and subscales. The resulting abuse scale
was normally distributed and internally consistent at all three assessment
points (s .89, .90, and .90, respectively).
Partner-reported victimization. Men's current or recent wives,
common-law partners, or girlfriends were asked to provide collateral
information on their partner's perpetration of abuse. To ensure comparability,
women reported victimization on the same 31-item abuse inventory
as was used for men's self-reported perpetration, with changes in wording
to indicate victimization rather than perpetration. Women's reports were
also grounded within an explicit 2-month time frame. Women's responses
to the 31-item scale were normally distributed and internally consistent
across all three assessment points (s .91, .84, and .85, respectively).
Responsibility for past abusive behavior. Perceptions of responsibility
for abusive behavior were assessed by asking men to rate their agreement
from 1 (not at all true) to 7 (very true) on three statements: one focused on
1 In general, Canadians conceptualize ethic background on the basis of
language and culture, yielding a broad range of cultural and ethnic groups.
Given the fact that the majority of this sample was Caucasian, we feel that
further breakdown of ethnicity is inappropriate.
2 Although obtaining partner information is a common difficulty in
research studies of batterers, clinical personnel usually have fewer difficulties.
The high rate of failed partner contact by the clinical program in
this study is not typical of this or other batterer programs. Instead, it
reflects personnel changes that occurred over the course of this study
resulting in the Battered Woman's Advocate position being vacant for a
period of approximately 3 months.
3 Stage-of-change ratings were not available at Time 1 for the remaining
15 men.
PATTERNS OF CHANGE AMONG BATTERERS 881
situational attributions of responsibility (i.e., "To understand my past
abusive behavior, you have to take a closer look at the situations surrounding
each abusive incident"), one focused on relational attributions of
responsibility (i.e., "To understand my past abusive behavior, you have to
take a closer look at my partner and the relationship that the two of us
have"), and the third focused on personal attributions of responsibility (i.e.,
"To understand my past abusive behavior, you have to take a closer look
at me"). Responsibility scores were derived by subtracting men's situational
and relational attributions from their personal attribution of responsibility
and adding a constant to move responses more toward the positive
range. Men's mean scores on the resulting index for three assessments were
M 5.96 (SD 3.36), M 7.72 (SD 4.26), and M 8.22 (SD 4.03).
Spouse-specific empathy. To assess level of empathy, we had men
complete the Self Dyadic Perspective-Taking Scale (Long, 1990). This
13-item instrument assesses men's strategies for perspective taking (8
items; e.g., "When I'm upset with my partner, I usually try to put myself
in her shoes for a while") and their empathetic awareness of their partner's
feelings, attitudes, and needs (5 items; e.g., "I very often seem to know how
my partner feels"), using a 5-point scale. The Self Dyadic Perspective-
Taking Scale has shown adequate validity and reliability (Long, 1990) and
was internally consistent in the current sample across assessment periods
(Perspective Taking, s .81, .78, and .84, respectively; Empathetic
Awareness, s .83, .87, and .73, respectively).
Spouse-specific communication skill. Men's communication skills
were assessed using a modified version of the Interpersonal Competence
Scale (Buhrmester, Furman, Wittenberg, & Reis, 1988). Men rated their
competence in a particular social skill area with a 5-point scale. Three of
the original five subscales (with eight items per subscale) were used in the
current study: Conflict Management (e.g., "refraining from saying things
that might cause a disagreement to build into a big fight"), Comfort With
Disclosure (e.g., "telling your partner things about yourself that you're
ashamed of"), and Emotional Supportiveness (e.g., "helping your partner
get to the heart of a problem that she is experiencing"). Each subscale has
been shown to be internally consistent and concurrently valid for young
adult peer relationships (Buhrmester et al., 1988) and adolescent dating
relationships (Wolfe, Wekerle, Scott, Straatman, & Grasley, 2002). Internal
consistency in the current sample was adequate across assessments
(Conflict Management, s .84, .90, and .87; Comfort With Disclosure,
s .81, .86, and .88; Emotional Supportiveness, s .87, .90, and .88).
Procedure
Program description. Participants were drawn from Changing Ways,
London. Changing Ways offers feminist-oriented group-based intervention
to voluntary and court-mandated batterers. The program philosophies,
intervention targets, and treatment style are consistent with North American
standards that guide service provision to this population of clients
(Dankwort & Austin, 1999). Men begin by attending a 3-hr meeting that
provides basic information about the nature of treatment and the expectations
of the agency. Men are then scheduled to attend an individual meeting
with a counselor to review their personal situation and ensure their eligibility
for the program.4 They then begin 10 weeks of open psychoeducational
counseling covering each section of the Minnesota Power and
Control Wheel (Pence & Paymar, 1986), which is intended to expand their
understanding of the range, function, and effects of abusive behavior.
Following these 10 sessions, they are invited to join a 7-week closed group.
During these sessions more attention is devoted to group process, insightoriented
interventions, and confrontations. Topics of these seven sessions
include men's accountability for their past abusive behavior, responsibility
for their current use of power, and the steps and strategies men need to
ensure the safety of their partners and children.
Research protocols. Men completed research questionnaires three
times over the course of treatment, beginning with their first meeting at the
agency. They were informed that their participation in the research was
voluntary and would in no way affect their status in the treatment program.
The second assessment occurred just prior to their last session in the
10-week group. On average, there were 17.6 weeks (SD 7.8) between
men's first and second assessment. The third assessment was completed
during their final group meeting, which occurred on average 10.5 weeks
(SD 3.3) after their second assessment. The questionnaires took approximately
1 hr to complete.
Research contacts with women were made following their initial clinical
contact with the Battered Women's Advocate, a staff member devoted to
providing referrals and support to men's partners. Women received information
about the ongoing study and indicated their consent or nonconsent
to be contacted by a researcher. Consenting women were then contacted by
telephone to provide a retrospective account of their partner's abuse
perpetration and responsibility during the 2 months prior to the time he
started treatment. Because of delays in scheduling clinical intakes, women's
initial research assessments were typically conducted 5 or 6 weeks
after their male partners had started treatment. Women were contacted for
second and third assessments as soon as possible after their partner had
completed their respective assessments. Neither men nor their partners
were paid for participating.
Analyses
Investigating longitudinal predictions of change requires statistical
methods capable of modeling both individual- and group-level change over
time. Fortunately, recent developments in covariance estimation for growth
curves (Bryk & Raudenbush, 1992) and the availability of computing
programs to carry out these analyses have made such methods available.
The use of these models offers many advantages to the study of change.
First, modeling techniques allow for the examination of various models of
growth (e.g., linear, quadratic) and for explicit tests of which change model
provides the best fit to the data. Second, because growth parameters rather
than Group
--------------------------------------------------------------------------------
Time interactions are outcome variables for person-level
analyses, these models can accommodate variance in the timing of observations
and a certain amount of missing individual-level data. Finally,
many of these models use an empirical Bayes's estimator, which maximizes
the reliability in estimation by weighing results according to the
precision of the relevant error variances.
For the current analyses, hierarchical linear modeling (Bryk & Raudenbush,
1992) was used to represent change through a two-level hierarchical
model, multiple observations nested within persons. Analyses proceeded in
two steps. First, the appropriate parameters to model within-subject growth
(i.e., change over time) for all potential outcome variables were specified.
Because at this step parameters are estimated without regard to any
potential moderating variables, the resulting models are labeled unconditional
growth models. Initial growth models for all outcome variables were
further specified through an examination of sociodemographic characteristics
(i.e., age, education, employment status, and arrest history), relationship
status (at treatment initiation and over the course of treatment), referral
source (i.e., court ordered or voluntary), and concurrent risk factors (i.e.,
childhood history of witnessed or experienced abuse, alcoholism) on men's
change over time. Controlling for significant effects of sociodemographic,
relationship, referral, and concurrent risk factors, in the next step we
determined whether stage of change had a significant effect on men's rate
of change in abusive behavior and in all other outcomes. Unless otherwise
specified, we estimated Level 1 variables (i.e., growth parameters) in all
analyses as random effects and Level 2 variables (i.e., factors expected to
affect growth) as fixed effects. Growth estimates were not centered, which
means that the intercept term represents men's status at the beginning of
treatment.
4 Men are typically refused service only if they have outstanding charges
for domestic assault or if they present a clear danger to other group
members.
882 SCOTT AND WOLFE
Results
Unconditional and Covariate-Adjusted Models for
Individual Growth
Unconditional growth was examined by modeling men's scores in
four outcome domains, abuse, empathy, responsibility, and communication,
over the three assessment periods. Because most outcome
domains were represented by two or more variables, the four domains
required eight independent analyses: two on indicators of abusive
behavior (self-reported perpetration and partner-reported victimization),
two on indicators of empathy (Empathetic Awareness and
Perspective Taking), three on indicators of communication skill (Conflict
Management, Comfort With Disclosure, and Emotional Supportiveness),
and the remaining one on the indicator of men's perception
of responsibility for past abusive behavior. Analysis of each of the
eight outcome variables began with an inspection of the pattern of
scores over time for individual men. Contrary to expectation, examination
of individual-level data revealed that growth tended to occur in
a distinctly nonlinear manner. For behaviors most directly associated
with abuse, the greatest changes in men's scores tended to occur
between the first and second assessment, followed by a period of
lesser change. Measures of skill tended to show change throughout
treatment, although often with differences in rate of change between
different points in time. These patterns of growth were modeled with
an intercept and two dummy-coded time parameters, one representing
the differences in men's scores (i.e., linear growth) from Assessments
1 to 2 and the other the difference in men's scores from
Assessments 2 to 3.5
To determine whether growth and variability in growth specified
by the unconditional model was dependent on men's characteristics,
we examined the effect of 10 potential covariates (age,
education, employment status, arrest history, referral source, relationship
status, change in relationship status, childhood history of
witnessed or experienced abuse, and alcoholism) on growth. All
potential covariates were initially entered into the growth equation
as fixed Level 2 effects, and nonsignificant covariates were eliminated
in a backward stepwise fashion. The growth parameters for
the resulting covariate-adjusted initial models are presented in
Table 1. The top section of this table presents the covariateadjusted
parameter estimates, t ratios, chi-square tests, and significance
values for the estimation of growth from Time 1 (T1) to
Time 2 (T2) for all eight outcome variables, whereas the bottom
half of the table presents the same information for growth from T2
5 Attempts to model this pattern of growth initially concentrated on a
model that included an intercept, a linear-growth parameter, and a
quadratic-growth parameter. Analysis of this model generally reduced to a
linear model that failed to adequately capture variations in men's change
over time. The dummy-coded model used in final analyses was created
following the suggestion of Bryk and Raudenbush (1992) for dealing with
data containing two distinct growth patterns. Although use of three growth
parameters (intercept and the two difference scores) to model change over
three assessments was a more accurate representation of the raw data, this
made it necessary to model change from Assessments 1 to 2 by assuming
fixed growth from Assessments 2 to 3 and to model change from Assessments
2 to 3 assuming fixed growth from Assessments 1 to 2.
Table 1
Covariate-Adjusted Models of Growth and Variability for Abusive Behavior, Responsibility,
Empathy, and Communication Skill
Dependent growth variable n B t-ratio 2
Time 1 to Time 2 parameter estimates
Self-reported abuse perpetrationa 89
--------------------------------------------------------------------------------
1.66
--------------------------------------------------------------------------------
4.55** 102.66**
Partner-reported abuse victimization 35
--------------------------------------------------------------------------------
1.60
--------------------------------------------------------------------------------
2.08* 282.80**
Responsibilitya 110 1.33 4.11** 102.63
Empathy
Empathetic awarenessb,c 102 0.15 2.53** 146.50**
Perspective takinga,b,f 105 0.39 1.19 154.45**
Communication
Conflict management 115 0.29 4.15** 123.42p.26
Comfort with disclosureb 94 0.17 2.20* 153.57**
Emotional supportiveness 115 0.18 2.93** 136.38p.07
Time 2 to Time 3 parameter estimates
Self-reported abuse perpetratione 106
--------------------------------------------------------------------------------
1.05
--------------------------------------------------------------------------------
2.94* 87.21
Partner-reported abuse victimization 89
--------------------------------------------------------------------------------
0.21
--------------------------------------------------------------------------------
0.71 19.26
Responsibility 110 0.44 1.36 100.57
Empathy
Empathetic awarenessd 102 0.08 1.77 77.20
Perspective taking 105 0.36 4.87** 92.44
Communication
Conflict management 115 0.35 4.95** 94.23
Comfort with disclosuree 94 0.21 3.67** 70.95
Emotional supportiveness 115 0.24 4.05** 101.92
Note. t ratio is a test of whether there is significant growth over time. Chi-square tests measure whether there
is significant variation in growth between individuals.
a Initial relationship status was used as analysis covariate. b Education was used as analysis covariate. c Age
was used as analysis covariate. d Alcoholism was used as analysis covariate. e Change in partner status over
treatment. f Arrest history was used as analysis covariate.
* p .05. ** p .01.
PATTERNS OF CHANGE AMONG BATTERERS 883
to Time 3 (T3). In addition, results indicate whether there was
significant growth over time (evaluated with a t ratio) and whether
there was significant variation in growth between individuals
(evaluated with chi-square tests).
Table 1 (top) shows that men had significant positive growth
across all but one outcome measure from T1 to T2. Men and their
partners both reported reductions in men's abusive behavior, and
men reported taking more responsibility for their abusive behavior.
Magnitude of change in men's reports of abuse was affected by
their initial relationship status, with greater reductions reported by
men with less than daily contact with their partners as compared
with men who had greater contact (i.e., were living with their partners
or having daily contact; B 2.34). Similarly, men who had little
contact with their partners showed greater increases in responsibility
than those with greater contact (B 1.15). Men also reported being
more aware of their partner's emotions over time, with older
(Bage 0.01) and more educated (Bdegree/diploma vs. other 0.07)
men showing slightly greater growth. Finally, men reported having
more skills for managing conflict, feeling more comfortable
with intimate disclosures, and providing emotional support to their
partners over time. Examined covariates contributed to prediction
only for comfort with disclosure, with more educated men reporting
greater gains in disclosure comfort (Bdegree/diploma vs. other
0.14). Examination of the chi-square test results shows significant
individual variation in men's change across most outcome
measures, including abusive behavior (both self- and partner reported),
empathetic awareness, perspective taking, and comfort
with disclosure. There was also substantial (but not statistically
significant) variability in growth for men's conflict management
and emotional supportiveness skill, but little or no interindividual
variability in growth in perceptions of responsibility.
Table 1 (bottom) indicates that positive growth tended to continue
between the second and third assessment periods for some
variables. Once again, men showed significant reductions in selfreported
abusive behavior, with men who separated from their
partners over treatment showing greater reductions than other men
(B 1.62). Men also reported gains in perspective taking, conflict
management, comfort with disclosure, and emotional supportiveness.
Of these changes, only perspective taking and comfort with disclosure
were significantly affected by covariates. Specifically, alcoholic men
reported less change in perspective taking than nonalcoholic and
possibly alcoholic men (B vs. nonalcoholic and possibly 0.36;
B vs. nonalcoholic 0.28), and men who separated from their partners
over treatment showed less change in disclosure comfort than men
who did not change partner status over treatment (B
--------------------------------------------------------------------------------
0.28).
Unfortunately, none of the latter changes had sufficient variability
to justify modeling Level 2 effects.
Stage of Change as a Predictor of Growth Over Time
The second step of analyses was to determine whether variability
in growth could be predicted by men's stage of change. Only
outcomes with sufficient interindividual variability (set liberally at
p .30) in the initial covariate-adjusted growth model were
examined; thus, analysis was restricted to change that occurred
from Assessments 1 to 2 in self- and partner-reported abuse,
empathy, and communication. Because stage of change is a categorical
variable, its effect was examined by adding three dummycoded
Level 2 fixed contrasts to the covariate-adjusted unconditional
models: men in the precontemplation versus contemplation
stage, precontemplation versus action stage, and contemplation
versus action stage. Contrast coefficients and their significance
values, presented in Table 2, reveal strong support for the utility of
stage of change as a predictor of variation in individual growth.
For five of the seven measures examined, growth varied significantly
by stage of change (i.e., at least one stage-of-change contrast
was significant). Exceptions were partner-reported victimization
and empathetic awareness, both of which approached
significance ( p .10). To aid in the interpretation of these results,
covariate-adjusted growth curves for all significant outcomes for
men in the precontemplation, contemplation, and action stages are
presented in Figures 1 through 4 and discussed in the following
sections.
Abusive behavior: Figures 1 and 2. These figures show selfand
partner-reported covariate-adjusted reductions in abusive behavior
for men in the precontemplation, contemplation, and action
Table 2
Parameter Estimates of the Effect of Stage of Change on Rate of Growth in Abusive Behavior,
Empathy, and Communication Skill From Time 1 to 2
Dependent growth
variable
Contrast 1: PC vs. C Contrast 2: PC vs. A Contrast 3: C vs. A
B t-ratio B t-ratio B t-ratio
Self-reported abuse
perpetrationa
--------------------------------------------------------------------------------
2.22
--------------------------------------------------------------------------------
2.59**
--------------------------------------------------------------------------------
1.66
--------------------------------------------------------------------------------
3.42**
--------------------------------------------------------------------------------
1.04
--------------------------------------------------------------------------------
1.08
Partner-reported abuse
victimization
--------------------------------------------------------------------------------
0.07
--------------------------------------------------------------------------------
0.10
--------------------------------------------------------------------------------
2.16
--------------------------------------------------------------------------------
1.81p.09
--------------------------------------------------------------------------------
4.23
--------------------------------------------------------------------------------
1.91p.07
Empathetic awarenessb,c 0.09 0.67 0.12 1.77p.08 0.12 0.76
Perspective takinga,b,d 0.26 1.29 0.32 2.72** 0.42 1.76p.08
Conflict management 0.38 2.54** 0.23 3.04** 0.09 0.48
Comfort with disclosureb 0.33 1.97* 0.24 2.87** 0.16 0.85
Emotional
supportiveness 0.16 1.15 0.17 2.40* 0.18 1.15
Note. PC precontemplation stage; C contemplation stage; A action stage.
a Initial relationship status was used as analysis covariate. b Education was used as analysis covariate. c Age
was used as analysis covariate. d Arrest history was used as analysis covariate.
* p .05. ** p .01.
884 SCOTT AND WOLFE
stages from Assessments 1 to 3. Consistent with our hypotheses,
men in the precontemplation stage showed, on average, virtually
no change in self- or partner-reported abusive behavior from
Assessment 1 to 2. In contrast, over the same period there was
substantial reduction in the self-reported abuse perpetration for
men in the contemplation and action stages (Figure 1) and in
partner-reported victimization for men in the action stage (Figure
2). Between Assessments 1 and 2, men in the contemplation stage
reduced their self-reported 2-month abuse incidence by 2.9 abusive
events. Men in the action stage reduced their self-reported abuse
incidence by almost 4 abusive events (3.7), and their partners
reported a 5-event reduction in their victimization. Interpretation
of stage-related differences in men's rate of change is complicated,
however, by the disparity of men's self-reported abuse at treatment
initiation. As a result of these initial differences, men's scores
became more rather than less similar over treatment (see Discussion
section).
Empathy: Figure 3. Examination of Figure 3 suggests that
men in the precontemplation and contemplation stages made slight
gains in their perspective-taking skills over the course of treatment:
0.2 and 0.4 points on a 5-point scale for precontemplative
and contemplative men, respectively. Men in the action stage, in
contrast, began treatment with particularly low scores in perspective
taking and made substantial "catch-up" gains over the first two
assessment periods (0.8-point change). Change in empathetic
awareness (not presented) showed the same stage-specific pattern
of results, although differences in rate of change failed to reach
significant levels.
Communication skill: Figure 4. Trajectories of change in disclosure,
emotional supportiveness, and conflict management skills
differed for men in the precontemplation, contemplation, and
action stages. Figure 4, which illustrates change trajectories in
men's comfort with disclosure, is shown as an example for all
three of these outcomes. Men in the precontemplation stage
showed, on average, virtually no gains in communication skill over
the first part of treatment. In contrast, men in the action stage and,
to a slightly lesser extent, men in the contemplation stage made
substantial changes in their self-reported communication skill from
Assessment 1 to Assessment 2: 0.3-, 0.2-, and 0.5-point gains,
respectively, on a 5-point scale for contemplative men and 0.5-,
0.4-, and 0.6-point gains, respectively, for active men on comfort
with disclosure, emotional supportiveness, and conflict resolution
skill. Once again, because of initial differences in the self-reported
communication skills of men in the precontemplation, contempla-
Figure 1. Growth curves over three assessment periods for men in the
precontemplation, contemplation, and action stages on self-reported abusive
behavior.
Figure 2. Growth curves over three assessment periods for men in the
precontemplation, contemplation, and action stages on partner-reported
victimization.
Figure 3. Growth curves over three assessment periods for men in the
precontemplation, contemplation, and action stages on empathetic
perspective-taking.
Figure 4. Growth curves over three assessment periods for men in the
precontemplation, contemplation, and action stages on comfort with disclosure.
PATTERNS OF CHANGE AMONG BATTERERS 885
tion, and action stages, men appeared more rather than less similar
as treatment progressed.
To gauge the clinical significance of changes made by men in
each of the stages of change over treatment, we compared rates of
physical abuse at pre- and posttreatment with rates found in the
general population. These results are not conceptually equivalent
to growth curve analyses and are presented for descriptive purposes
only. At treatment initiation, 35% of program participants
reported at least one incident of physical abuse perpetration. This
can be translated to pretreatment 2-month incidence rates of physical
abuse for men in the precontemplation (352 per 1,000), contemplation
(355 per 1,000), and action (381 per 1,000) stages; all
are well above the 113 per 1,000 annual incidence rate reported in
the NFVS (Straus & Gelles, 1986). Over the course of treatment,
13.4% of men reported perpetrating at least one incidence of
physical abuse, with stage-related differences. Specifically, 70% of
reassaults were perpetrated by men in the precontemplation stage,
20% by men in the contemplation stage, and 10% by men in the
action stage. These stage differences approached, but failed to
reach, significant levels, 2(2, N 86) 5.24, p .07. Conversion
to 2-month incidence yields rates of 190 per 1,000 for men in
the precontemplation stage, which remains higher than the annual
incidence rate reported in the NFVS, as compared with 36 and 52
per 1,000 for men in the contemplation and action stages, respectively.
Although promising as an indicator of success for men in
the contemplation and action stages, it is important to remember
that these rates represent 2-month, rather than annual, incidence. If
these rates of abuse are directly multiplied to estimate annual
incidence, physical abuse incidence for men in all groups remains
higher than national averages.
Discussion
This study examined whether variations in change among men
completing a batterer intervention program could be predicted
from the TTM's stages of change. Results provide some evidence
that men who are in the contemplation and action stages at the
beginning of treatment show substantially greater positive growth
across a range of outcome measures than men who are in the
precontemplation stage. Specifically, stage-related differences in
growth were noted for self- and partner-reported abusive behavior
and for self-reported perspective taking, conflict management,
disclosure, and emotional support skills from the beginning to the
midpoint in treatment. Over the latter part of treatment, estimates
of growth tended to lack sufficient variability to model individuallevel
differences; thus, hypothesized effects of stage of change
could not be reliably examined.
The lack of change in precontemplative men, as compared with
men in the contemplation and action stages, is consistent with
numerous other studies applying the TTM to a wide range of
behavior problems (Pallonen et al., 1994). These findings are of
particular note when considered in light of past research on change
in abusive behavior that has identified few clinically and empirically
meaningful strategies for predicting which men will benefit
from treatment (Gondolf, 2002). Despite this apparent promise of
the utility of the TTM for batterers, significant issues remain. Most
important, despite stage-related differences in rate of change, men
appear to become more rather than less similar over treatment.
This apparent contradiction occurs because men in the precontemplation
stage reported less abuse and greater communication and
empathy skills at treatment initiation than men in later stages.
Interpretation of the importance of stage of change therefore
depends on whether greater weight is placed on amount of change
over time, which varies by stage of change, or status at final
assessment, which appears to be similar for men across stages.
Consideration of response biases is important for clarifying
these different perspectives on outcome. Although social desirability
and impression management are inconsistently related to selfreports
of abusive behavior in past studies (Dutton & Hemphill,
1992; Sugarman & Hotaling, 1997), there is now fairly consistent
evidence that stage of change marks differential biases in the
reporting of abusive men, with men in the precontemplation stage
showing the greatest bias (Levesque et al., 2000; Murphy et al.,
1999). On the basis of the current sample of men, we also found
that, despite self-reported relationship health, partners of men in
the precontemplation stage at treatment initiation reported as much
victimization as partners of men in later stages (Scott & Wolfe,
2002). Precontemplative men were also as likely as others to have
had their relationships end and to have been arrested. Thus, there
are good reasons to suspect that pretreatment differences across
stages are related to denial, as suggested by the TTM, and are not
real differences in abusiveness, empathy, or communication skill.
By locating a major source of self-report bias with precontemplative
men, we draw attention to the potential misinterpretation of
poor outcomes among this subsample. Because of the confound of
self-reporting biases, change over time may reflect changes in
honesty and openness as much as it does changes in abusive
behavior, and any pattern of progress is potentially indicative of
positive change. For example, positive change for this subgroup of
clients could be indicated by increased reports of abuse (because of
greater recognition or willingness to report), stable reports of abuse
(greater recognition followed by reduction in behavior), or even
reductions in abuse reports. A detailed examination of change
among individual men in the precontemplation stage suggests that
all of these changes are occurring, and to a much greater extent
than for men in the contemplation and action stages. Although the
vast majority of men in the contemplation and action stages
reported overall reductions in abusive behavior over treatment,
response patterns of men in the precontemplation stage were more
varied, with approximately one third reporting increases in abuse,
one third stability, and one third reductions.
Despite this caveat, there are a number of reasons from the
current study to suspect that men in the precontemplation stage do
indeed make less change over treatment then men in the contemplation
and action stages. First, localization of reporting biases
primarily with precontemplative men allows for somewhat greater
confidence in positive results for men in the contemplation and
action stages of change. Men in the contemplation and action
stages did report greater benefits from intervention on a broad
range of intervention measures than men in the precontemplation
stage. Second, although based on a limited sample, women's
reports of victimization did show differences in the abuse of men
in different stages at treatment termination. Specifically, partners
of men in the action stage reported greater rates of change in
victimization and had posttreatment scores that were approximately
50% lower than those reported by partners of men in the
contemplation and precontemplation stages. Interestingly, qualitative
studies of women's perspectives of their partner's progress
886 SCOTT AND WOLFE
through treatment also emphasize the importance of men's readiness
to change and problem denial. Specifically, women who feel
that treatment did not work for their partners tend to attribute this
failure to their partners' denial of his abusiveness and unwillingness
to make any effort to change (Gregory & Erez, 2002)-both
characteristics of the precontemplation stage.
Third, current results show that when only physical abuse is
considered, there are clinically relevant, stage-related differences
in abusive behavior at treatment termination. At final assessment,
rates of physical abuse for men in the precontemplation stage are
higher than those of men in the contemplation or action stages and
higher than national incidence rates. It is suspected that this difference
occurs because physically abusive events, such as a slap or
a punch, are less open to distortion and minimization than the more
common emotionally abusive events, such as isolating or degrading
a partner. These events may, therefore, provide an estimate of
abusiveness that is less affected by men's denial.
In summary, the stage-related differences in men's rates of
change over treatment likely reflect actual differences in men's
progress, despite apparent similarity in men's scores at final assessment.
However, considerably more research is needed to verify
this conclusion and to determine how best to partial out the
effects of reduced denial from the change progress. At least a
subset of precontemplative men may benefit from treatment. Subsequent
studies may wish to rely to a greater extent on objective
rather than subjective reports of change. Observational measures,
though time consuming, may provide a better assessment of the
difficulties in men's self-reported and actual skills and attitudes.
Alternatively, clinically based measures that include social desirability
and denial scales along with a range of behaviors and
attitudes associated with abusive behavior may be useful (e.g., the
Personal and Relationship Scale; Straus, Hamby, Boney-McCoy,
& Sugarman, 1999).
A second question raised by the current results concerns the
validity of the distinction between abusive men in the contemplation
and action stages. Differences in rate of change for these
groups of men approached significance on partner-reported victimization.
However, on self-reported outcomes, the rates of
change for contemplative and active men tended to be quite
similar. It is interesting to speculate on whether these findings are
a function of measurement error, source of data (i.e., self- or
partner-report), or sample bias or whether the distinction between
contemplation and action has no predictive validity with this
population. It is also possible that the use of generic "problem"
wording, rather than explicit mention of intimate partner abuse,
limited our ability to differentiate men according to change in
abusive behavior. Alternatively, stage categorizations may need to
be abandoned in favor of continuous measures of dimensions that
underlie readiness to change, such as problem denial, change
motivation, and change efficacy; or men may need to be classified
according to progress through the stages of change over time (i.e.,
change in stage of change).
Results of the current study also raise issues for the broader
study of batterer treatment. A strength of the current study is its use
of statistical methods explicitly designed to model individual
change. The use of these methods provided statistical confirmation
that there is a great deal of variability in men's change over time.
In addition, results draw attention to the need to consider patterns
of growth that, in the current study, varied depending on the
specific outcome measure being examined. On average, men made
changes in empathy, perceptions of responsibility, and abusive
behavior early in treatment, whereas improvements in men's communication
skill showed more continuous growth over time. These
results imply that judgments of the success of a treatment program
may depend on what measures are being examined over which
time periods. If abusive behavior is the only outcome being examined,
as is often the case, shorter programs may show similar
outcomes to longer ones. However, if a broader range of outcomes
is sought, then longer intervention may be necessary.
Although the current study has many strengths, its results must
be considered in light of its limitations. First, the low participation
rate of men's partners limits the confidence that can be placed in
interpretation results. This is particularly the case for men in the
contemplation stage, where results obtained with data from selfand
partner-report differed considerably. Moreover, the reports of
both men and women may be biased by the short time frame used
in reporting intimate abuse (i.e., 2 months), with particular concern
that clients' frame of reference at the first assessment was greater
than 2 months. To mitigate against this, clients were directed to
identify an event that occurred 2 months earlier as a starting point
for their reporting; however, this strategy may not have corrected
sufficiently for recall biases. A second limitation concerns data
points. Because only three data points were available and a linear
growth model was inappropriate, it was necessary to fix one
parameter in all analyses. Although fixing a parameter was theoretically
and empirically justifiable, results would be strengthened
by the addition of a fourth data point. Third, the current study
concentrated only on men in treatment. Subsequent studies would
benefit from the inclusion of waiting list or dropout control groups
and a longitudinal follow-up. It would also be informative to
examine men's personality and posttreatment alcohol use as potential
additive or alternate predictors of outcome. Finally, as
previously mentioned, the conclusions of the current study are
limited by the lack of an appropriate index of men's denial.
Despite limitations, the results of this study raise interesting
possibilities for future research and provide hopeful leads for
identifying men who are unlikely to benefit from batterer treatment.
It is recommended that future research clarify the interface
between change readiness and treatment outcome, with the specific
aim of determining whether precontemplative men can benefit
from standard treatment. The TTM suggests that, for precontemplative
men, treatments focused on consciousness raising, dramatic
relief, and environmental reevaluation are the most helpful
(Prochaska et al., 1992). Motivational interviewing may also be of
particular benefit to this subgroup of clients (Miller & Rollnick,
2002). On the other hand, coercing abusive men into standard
treatments not specifically targeting their needs may be iatrogenic,
particularly if men blame women for their need to attend a program
(Rosenfeld, 1992). Even if abuse rates among precontemplative
men do not actually worsen during treatment, a trend for them
to stay at relatively stable levels must be considered in light of the
effects of treatment on men's partners. As Gondolf (1988) demonstrated
over 10 years ago, men's attendance at a treatment
program greatly increases the chance that their partners will return
to a relationship with them. If subsequent studies confirm the
predictive value of men's stage of change, it is ethically necessary
to ask whether current treatment models need to be modified
significantly to address the needs of this population of men.
PATTERNS OF CHANGE AMONG BATTERERS 887
References
Adams, D. (1985). Stages of anti-sexist awareness and change for men who
batter. In L. J. Dickstein & C. C. Nadelson (Eds.), Family violence:
Emerging issues of a national crisis (pp. 61-98). Washington, DC:
American Psychiatric Press.
Austin, J. B., & Dankwort, J. (1999). Standards for batterer programs: A
review and analysis. Journal of Interpersonal Violence, 14, 152-168.
Begun, A. L., Murphy, C. M., Bolt, D., Weinstein, B., Strodthoff, T., Short,
L., & Shelley, G. (2003). Characteristics of the Safe at Home instrument
for assessing readiness to change intimate partner violence. Research on
Social Work Practice, 13, 80-107.
Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models:
Advanced quantitative techniques in the social sciences series. Newbury
Park, CA: Sage.
Buhrmester, D., Furman, W., Wittenberg, M. T., & Reis, H. T. (1988). Five
domains of interpersonal competence in peer relationships. Journal of
Personality and Social Psychology, 55, 991-1008.
Daly, J. E., & Pelowski, S. (2000). Predictors of dropout among men who
batter: A review of studies with implications for research and practice.
Violence and Victims, 15, 137-160.
Daniels, J. W., & Murphy, C. M. (1997). Stages and processes of change
in batterers' treatment. Cognitive and Behavioral Practice, 4, 123-145.
Dankwort, J., & Austin, J. (1999). Standards for batterer intervention
programs in Canada: A history and review. Canadian Journal of Community
Mental Health, 18, 19-38.
Dunford, F. W. (2000). The San Diego Navy experiment: An assessment of
interventions for men who assault their wives. Journal of Consulting and
Clinical Psychology, 68, 468-476.
Dutton, D. G. (1995). The domestic assault of women. Vancouver, British
Columbia, Canada: UBC Press.
Dutton, D. G., Bodnarchuk, M., Kropp, R., Hart, S. D., & Ogloff, J. P.
(1997). Client personality disorders affecting wife assault post-treatment
recidivism. Violence & Victims, 12, 37-50.
Dutton, D. G., & Hemphill, K. J. (1992). Patterns of socially desirable
responding among perpetrators and victims of wife assault. Violence &
Victims, 7, 29-39.
Fals-Stewart, W., Kashdan, T. B., O'Farrell, T. J., & Birchler, G. R. (2002).
Behavioral couples therapy for drug-abusing patients: Effects on partner
violence. Journal of Substance Abuse Treatment, 22, 87-96.
Faulkner, K., Stoltenberg, C. D., Cogen, R., Nolder, M., & Shooter, E.
(1992). Cognitive-behavioral group treatment for male spouse abusers.
Journal of Family Violence, 7, 37-55.
Gondolf, E. W. (1987). Changing men who batter: A developmental model
for integrated interventions. Journal of Family Violence, 2, 335-349.
Gondolf, E. W. (1988). The effect of batterer counseling on shelter outcome.
Journal of Interpersonal Violence, 3, 275-289.
Gondolf, E. W. (1997). Batterer programs: What we know and need to
know. Journal of Interpersonal Violence, 12, 83-98.
Gondolf, E. W. (2002). Batterer intervention systems: Issues, outcomes,
and recommendations. Thousand Oaks, CA: Sage.
Gregory, C., & Erez, E. (2002). The effects of batterer intervention programs:
The battered woman's perspective. Violence Against Women, 8,
206-232.
Hamberger, L. K., & Hastings, J. E. (1988). Skills training for treatment of
spouse abusers: An outcome study. Journal of Family Violence, 3,
121-130.
Hamberger, L. K., & Hastings, J. E. (1990). Recidivism following spouse
abuse abatement counseling: Treatment program implications. Violence
& Victims, 5, 157-170.
Johnson, S. M., Hunsley, J., Greenberg, L., & Schindler, D. (1999).
Emotionally focused couples therapy: Status and challenges. Clinical
Psychology: Science & Practice, 6, 67-79.
Kirsch, I. (1999). How expectancies shape experience. Washington, DC:
American Psychological Association.
Levesque, D., Gelles, R., & Velicer, W. (2000). Development and validation
of a stage of change measure for men in batterer treatment. Cognitive
Therapy and Research, 24, 175-199.
Long, E. C. J. (1990). Measuring dyadic perspective-taking: Two scales for
assessing perspective-taking in marriage and similar dyads. Educational
and Psychological Measurement, 50, 91-103.
McConnaughy, E. A., DiClemente, C. C., Prochaska, J. O., & Velicer,
W. F. (1989). Stages of change in psychotherapy: A follow-up report.
Psychotherapy, 26, 494-503.
Miller, W. R., & Rollnick, S. (2002). Motivational interviewing: Preparing
people for change (2nd ed.). New York: Guilford Press.
Murphy, C. M., Begun, A., & Strodthoff, T. (1999, November). Concurrent
and predictive validity of an instrument to assess readiness to
change abusive behavior toward intimate partners. Poster presented at
the 33rd Annual Meeting of the Association for Advancement of Behavior
Therapy, Toronto, Ontario, Canada.
Murphy, C. M., Musser, P. H., & Maton, K. I. (1998). Coordinated
community intervention for domestic abusers: Intervention system involvement
and criminal recidivism. Journal of Family Violence, 13,
263-284.
National Clearinghouse on Family Violence. (2002). Canada's treatment
programs for men who abuse their partners. Ottawa, Ontario, Canada:
Minister of Health Canada.
Ockene, J., Kristeller, J. L., Goldberg, R., Ockene, I., Merriam, P., Barrett,
S., et al. (1992). Smoking cessation and severity of disease: The Coronary
Artery Smoking Intervention Study. Health Psychology, 11, 119-
126.
Pallonen, U. E., Leskinen, L., Prochaska, J. O., Willey, C. J., Kaarianen,
R., & Salonen, J. T. (1994). A 2-year self-help smoking cessation
manual intervention among middle-aged Finnish men: An application of
the transtheoretical model. Preventive Medicine, 23, 507-514.
Pence, E., & Paymar, M. (1986). Power and control: Tactics of men who
batter. Duluth: Minnesota Program Development.
Petrik, N. D., Gildersleeve-High, L., McEllistrem, J. E., & Subotnik, L. S.
(1994). The reduction of male abusiveness as a result of treatment:
Reality or myth? Journal of Family Violence, 9, 307-316.
Prochaska, J. O., DiClemente, C. C., & Norcross, J. C. (1992). In search of
how people change: Applications to addictive behaviors. American
Psychologist, 47, 1102-1114.
Prochaska, J. O., Velicer, W. F., Rossi, J. S., Goldstein, M. G., Marcus,
B. H., Rakowski, W., et al. (1994). Stages of change and decisional
balance for 12 problem behaviors. Health Psychology, 13, 39-46.
Roberts, A. (Ed.). (1984). Intervention with the abusive partner. In Battered
women and their families (pp. 84-115). New York: Springer.
Rosenfeld, B. D. (1992). Court-ordered treatment of spouse abuse. Clinical
Psychology Review, 12, 205-226.
Saunders, D. G. (1996). Feminist-cognitive-behavioral and process-
psychodynamic treatments for men who batter: Interaction of abuser
traits and treatment models. Violence and Victims, 11, 393-413.
Scalia, J. (1994). Psychoanalytic insights and the prevention of pseudosuccess
in the cognitive-behavioral treatment of batterers. Journal of Interpersonal
Violence, 9, 548-555.
Scott, K. L. (2002). Predictors of change among batterers: Application of
theories and review of empirical findings. Manuscript submitted for
publication.
Scott, K. L., & Wolfe, D. A. (2000). Change among batterers: Examining
men's success stories. Journal of Interpersonal Violence, 15, 827-842.
Scott, K. L., & Wolfe, D. A. (2002). Application of the transtheoretical
model's stages of change to abusive men. Manuscript submitted for
publication.
Scott, K. L., & Wolfe, D. A. (in press). Stage of change as a predictor of
attrition among men in a batterer treatment program. Journal of Family
Violence.
Selzer, M. L., Vinokur, A., & van Rooijen, L. (1975). A self-administered
888 SCOTT AND WOLFE
Short Michigan Alcoholism Screening Test (SMAST). Journal of Studies
on Alcohol, 36, 117-126.
Straus, M. A. (1979). Measuring intrafamily conflict and violence: The
Conflict Tactics (CT) Scales. Journal of Marriage and the Family, 41,
75-88.
Straus, M. A., & Gelles, R. (1986). Societal change and change in family
violence as revealed by two national surveys. Journal of Marriage and
the Family, 48, 465-479.
Straus, M. A., & Gelles, R. (1990). Physical violence in American families.
New Brunswick, NJ: Transaction.
Straus, M. A., Hamby, S. L., Boney-McCoy, S., & Sugarman, D. (1999).
The Personal and Relationships Profile (PRP). Durham: Family Research
Laboratory, University of New Hampshire.
Sugarman, D. B., & Hotaling, G. T. (1997). Intimate violence and social
desirability: A meta-analytic review. Journal of Interpersonal Violence,
12, 275-290.
Sutton, S. (2001). Back to the drawing board? A review of applications of
the transtheoretical model to substance use. Addiction, 96, 175-196.
Taft, C. T., Murphy, C. M., King, S. W., Musser, P. C., & DeDeyn, J. M.
(2003). Process and treatment adherence factors in group cognitive-
behavioral therapy for partner-violent men. Journal of Consulting and
Clinical Psychology, 71, 812-820.
Tolman, R. M. (1999). The validation of the Psychological Maltreatment of
Women Inventory. Violence and Victims, 4, 173-189.
Weinstein, N. D., Rothman, A. J., & Sutton, S. R. (1998). Stage theories
and health behavior: Conceptual and methodological issues. Health
Psychology, 17, 290-299.
Wolfe, D. A., Wekerle, C., Scott, K. L., Straatman, A., & Grasley, C.
(2002). Comparative pathways to dating violence for adolescent
girls and boys: Prediction over one year. Manuscript submitted for
publication.
Received March 28, 2002
Revision received April 4, 2003
Accepted April 21, 2003
PATTERNS OF CHANGE AMONG BATTERERS 889
Jessie Rayl
EM: thedogmom63 at frontier.com
PH:304.671.9780
www.facebook.com/eaglewings10
"But they that wait upon the LORD shall renew their strength; they shall
mount up with wings as eagles. They shall run, and not be weary"--Isaiah 40.31
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.acb.org/pipermail/acb-hsp/attachments/20110725/36478502/attachment-0001.html>
More information about the acb-hsp
mailing list