Assessing Relationship Quality in Mandated Community Treatment: Blending Care With Control

Department of Psychology and Social Behavior, University of California, Irvine, CA 92697-7083, USA.
Psychological Assessment (Impact Factor: 2.99). 01/2008; 19(4):397-410. DOI: 10.1037/1040-3590.19.4.397
Source: PubMed
Traditional measures of the therapeutic alliance do not capture the dual roles inherent in relationships with involuntary clients. Providers not only care for, but also have control over, involuntary clients. In 2 studies of probationers mandated to psychiatric treatment (n=90; n=322), the authors developed and validated the revised Dual-Role Relationships Inventory (DRI-R). The authors found that (a) relationship quality in mandated treatment involves caring and fairness, trust, and an authoritative (not authoritarian) style, (b) the DRI-R assesses these domains of relationship quality, is internally consistent, and relates in a theoretically coherent pattern with ratings of within-session behavior and with measures of the therapeutic alliance, relationship satisfaction, symptoms, and treatment motivation, and (c) the quality of dual-role relationships predicts future compliance with the rules, as assessed by probation violations and revocation. The DRI-R covaries with multiple domains more strongly than a leading measure of the therapeutic alliance, suggesting that it better captures the nature and effect of relationship quality in mandated treatment.


Available from: Devon L L Polaschek
Assessing Relationship Quality in Mandated Community Treatment:
Blending Care With Control
Jennifer L. Skeem and Jennifer Eno Louden
University of California, Irvine
Devon Polaschek
Victoria University of Wellington
Jacqueline Camp
University of Nevada, Las Vegas
Traditional measures of the therapeutic alliance do not capture the dual roles inherent in relationships
with involuntary clients. Providers not only care for, but also have control over, involuntary clients. In
2 studies of probationers mandated to psychiatric treatment (n 90; n 322), the authors developed and
validated the revised Dual-Role Relationships Inventory (DRI–R). The authors found that (a) relationship
quality in mandated treatment involves caring and fairness, trust, and an authoritative (not authoritarian)
style, (b) the DRI–R assesses these domains of relationship quality, is internally consistent, and relates
in a theoretically coherent pattern with ratings of within-session behavior and with measures of the
therapeutic alliance, relationship satisfaction, symptoms, and treatment motivation, and (c) the quality of
dual-role relationships predicts future compliance with the rules, as assessed by probation violations and
revocation. The DRI–R covaries with multiple domains more strongly than a leading measure of the
therapeutic alliance, suggesting that it better captures the nature and effect of relationship quality in
mandated treatment.
Keywords: therapeutic alliance, dual-role relationships, mandated treatment, probation, mental health
The quality of the relationship between a service provider
and a client is a “quintessential integrative variable” that cuts
across different modes of treatment (Wolfe & Goldfried, 1988,
p. 449). The provider– client relationship affects patient satis-
faction, treatment adherence, and outcomes in psychotherapy
(Krupnick et al., 1996), psychiatric treatment (Alverson, Alver-
son, & Drake, 2000; Cruz & Pincus, 2002; Frank & Gunderson,
1990; Howgego, Yellowlees, Owen, Meldrum, & Dark, 2003;
McCabe & Priebe, 2004; Penn et al., 2004; Solomon, Draine, &
Marcus, 2002), substance abuse treatment (Connors, Carroll,
DiClemente, Longabaugh, & Donovan, 1997), medical care
(e.g., Cooper-Patrick et al., 1999; Hall, Horgan, Stein, & Roter,
2002; Kaplan et al., 1996; Kaplan, Greenfield, & Ware, 1989),
and interventions for criminal behavior (Brown & O’Leary,
2000; Taft, Murphy, King, Musser, & DeDeyn, 2003). The size
of the effect is larger than that of other processes that can be
influenced by a provider. For example, the quality of the
treatment relationship in psychotherapy shapes patient out-
comes more strongly than the specific techniques applied (Asay
& Lambert, 1999; Horvath & Symonds, 1991; Luborsky et al.,
2002; Martin, Garske, & Davis, 2000). Moreover, the outcomes
affected are diverse, ranging from improvement in symptoms
for psychiatric patients (Cruz & Pincus, 2002) to a reduced risk
of violence for criminal offenders (Brown & O’Leary, 2000).
Such findings suggest that it would be a mistake to define
evidence-based practices solely as technical procedures, with-
out including the process of care. Across forms of treatment,
communication “is the main ingredient in . . . care and it is the
fundamental instrument by which therapeutic goals are
achieved” (Roter & Hall, 1992, p. 3).
In most studies of relationship quality and treatment out-
comes, investigators have simply imported measures developed
in conventional psychotherapy contexts to assess the therapeu-
tic alliance. These measures may not capture relationship qual-
ity in the context of mandated treatment. When patients are
required to take part in treatment, providers have dual roles:
They not only care for, but also have control over, the patient.
In the present study, we developed and validated a measure of
relationship quality in mandated treatment. We begin by out-
lining the need for such a measure.
Jennifer L. Skeem and Jennifer Eno Louden, Department of Psychology
and Social Behavior, University of California, Irvine; Devon Polaschek,
School of Psychology, Victoria University of Wellington, Wellington, New
Zealand; Jacqueline Camp, Department of Psychology, University of Ne-
vada, Las Vegas.
This research was supported by a grant from the MacArthur Research
Network on Mandated Community Treatment. We thank Paula Emke-
Francis for her work in coordinating data collection; the success of this
project is a product of her efforts. We also thank Susan Stodola and Steven
Lessard for kindly facilitating access to the probation team; Jeff Grobe and
Breanne Carmack for interviewing participants; executives of the Ameri-
can Parole and Probation Association, including Carl Wicklund, the Na-
tional Association of Probation Executives, including Melissa Cahill, and
the Council of State Governments, including Michael Thompson, for
supporting this research.
Correspondence concerning this article should be addressed to Jennifer
L. Skeem, Department of Psychology and Social Behavior, University of
California, Irvine, 3311 Social Ecology II, Irvine, CA 92697-7083. E-mail:
Psychological Assessment Copyright 2007 by the American Psychological Association
2007, Vol. 19, No. 4, 397–410 1040-3590/07/$12.00 DOI: 10.1037/1040-3590.19.4.397
Page 1
Treatment Mandates and Pressures Are Common in
Contemporary Community Treatment
As observed by Howgego et al. (2003), conventional measures
of the therapeutic alliance may poorly fit patients who “do not
voluntarily seek help and enter a relationship motivated to engage”
(p. 180). Such patients may compose a large group. First, a variety
of legal tools are now being used to require that patients attend
treatment and take psychotropic medication. Formal treatment
mandates come in many forms, including involuntary outpatient
commitment and special conditions of probation, which are civil
and criminal judicial orders (respectively) for a patient to adhere to
a community treatment plan. Based on a sample of 1,000 outpa-
tients drawn from public community outpatient settings, Monahan
et al. (2005) found that nearly half (44%– 66%) had experienced at
least one of four types of formal mandates to participate in treat-
Second, patients are subject to informal pressure from others to
adhere to treatment. Patients often have case managers who can be
charged with keeping patients in treatment (Hellerstein, Rosenthal,
& Miner, 1995), sometimes through assertive outreach efforts that
involve taking treatment to the patient in the community, whether
the patient wants it or not (McCabe & Priebe, 2004). Case man-
agers use a variety of strategies to pressure clients to engage in
treatment and change their behavior (Angell & Mahoney, 2007;
Angell, Mahoney, & Martinez, 2006). Based on a sample of 1,564
veterans who had been treated by assertive case management
teams, Neale and Rosenheck (2000) found that case managers
routinely used strong verbal guidance (e.g., reminding a client to
do or not do certain things) and often used money management to
control behavior. Less often, they used contingent withholding of
help, hospitalization, and appeals to external authorities. Conven-
tional measures of the therapeutic alliance do not capture the social
control inherent in these relationships.
Dual-Role Relationships Are More Complex Than the
Therapeutic Alliance
Despite the routine use of mandates and pressures in treatment,
little is known about how relationship quality may be altered in
this context. There are both conceptual and practical reasons for
developing a measure of relationship quality in mandated treat-
ment. Conceptually, it is difficult to reduce relationship quality
with involuntary patients to traditional notions of the therapeutic
alliance. First, although there are several conceptualizations of this
construct, most measures of the alliance tap (a) an affective bond
or attachment and (b) collaboration or willingness to invest in the
therapy process (Henry & Strupp, 1994; Horvath & Luborsky,
1993). When treatment is mandated, true collaboration and part-
nership may be lacking. The provider’s control over the patient
seems to render the alliance lopsided. Second, in mandated treat-
ment, one has dual roles. Reconciling one’s “helping, therapeutic,
or problem-solving role” with one’s controlling or “surveillance
role” (Trotter, 1999) may be both the most difficult and most
important component of effective work with involuntary patients
(D. A. Andrews, Zinger, Hoge, & Bonta, 1996; Klockars, 1972;
Trotter, 1999). Traditional measures of the alliance focus narrowly
on the therapeutic part of this relationship, even though the manner
in which the controlling part of the relationship is implemented
may become an integral component of relationship quality as a
Beyond the poor conceptual fit of the therapeutic alliance to
dual-role relationships, there are practical reasons for developing a
measure of relationship quality for the context of mandated treat-
ment. The goals of mandated treatment are more complex than
those in voluntary care. With patients who are mandated to treat-
ment, one is interested in achieving both compliance with treat-
ment and other requirements (i.e., rule compliance) and positive
traditional clinical outcomes. Measures that capture both the caring
and controlling aspects of relationship quality may better predict
these multifaceted outcomes than traditional measures of the alli-
ance. Indeed, effectively reconciling dual roles to establish a
trusting relationship with involuntary patients may be pivotal in
engaging them in treatment and achieving positive outcomes
(Drake, Wallach, Alverson, & Mueser, 2002).
The components of relationship quality in mandated treatment
are largely undefined. The first steps toward defining and assessing
relationship quality in mandated treatment involve (a) selecting an
appropriate context for studying these relationships and (b) iden-
tifying the relationships’ likely contours.
Choosing an Ideal Context for Measure Development
Patients with co-occurring mental and substance abuse disorders
are an appropriate population for studying dual-role relationships,
given that treatment mandates and pressures are disproportionately
applied to these patients. Relative to patients without co-occurring
disorders, those with co-occurring disorders are at double the risk
of having formal treatment conditions imposed on them (Monahan
et al., 2005). Given that patients with co-occurring disorders are
grossly overrepresented in correctional settings (Abram & Teplin,
1991), treatment is often mandated as a function of patients’
involvement in the criminal justice system. The vast majority of
individuals arrested are placed on probation (Bureau of Justice
Statistics, 2006), and those with mental disorders typically are
required to participate in treatment as a special condition of pro-
bation (Ditton, 1999; U.S. Probation and Pretrial Services, 2001).
A similar picture emerges with informal pressure to participate
in treatment. Case managers are likely to use treatment pressures
with patients who have severe symptoms, recent drug use, and
arrest histories (Neale & Rosenheck, 2000). Similarly, case man-
agers for probationers often fall prey to the “treater-turned-
monitor” phenomenon, chiefly monitoring for treatment noncom-
pliance and elevating probationers’ risk of incarceration on a
technical violation (Solomon et al., 2002). For patients with the
triple stigma (Hartwell, 2004) of mental disorder, substance abuse,
and criminal justice involvement, treatment relationships are often
infused with social control.
For these reasons, an ideal context for developing a measure of
relationship quality in mandated treatment is specialty mental
health probation programs. This was the context in which the
Page 2
present study was conducted. In specialty programs, officers with
mental health training supervise reduced caseloads composed
solely of probationers who predominantly have co-occurring men-
tal and substance abuse disorders (Skeem, Emke-Francis, & Eno
Louden, 2006). These officers function much like case managers,
in that they advocate for social services (e.g., psychiatric treat-
ment, Social Security Disability Income, housing), coordinate
closely with providers, and work directly with clients toward
therapeutic goals (Skeem et al., 2006). Indeed, the majority of
officers in the present study had once been case managers. Unlike
traditional case managers, however, these officers are explicitly
tasked with managing dual roles that place equal weight on care
(rehabilitation) and control (rule compliance and public safety).
These officers are experienced not only in advocating for services
but also in monitoring and enforcing the conditions of probation,
including the special condition to take psychotropic medication
and participate in treatment. They are a logical choice, given that
they provide care to and implement treatment mandates for a
growing high-risk population.
Identifying Likely Contours of Relationship Quality
A small literature provides a glimpse of the importance and
nature of relationship quality in the context of mandated treatment.
In a multisite focus group study, Skeem, Encandela, and Eno
Louden (2003) found that probationers with mental disorder and
their officers believed that the quality of their relationships colored
every interaction and strongly influenced clinical and criminal
outcomes. Harmful relationships were described as authoritarian
ones characterized by many demands, little flexibility, and belit-
tling use of control. These relationships were perceived as ongoing
stressors that compromised probationers’ mental state and func-
tioning and sometimes engendered reactance to officers’ direc-
tives. In helpful relationships, the affiliative aspects of the thera-
peutic alliance were blended with social control. Here, however,
control was used in the right way, that is, in a manner perceived as
fair, respectful, and motivated by caring. This manner may be
viewed as an explicitly interpersonal form of procedural justice
(see MacCoun, 2005). Such relationships provided support, en-
couraged trust, and instilled a desire to please officers.
The concept of procedural justice has also proven crucial to
providers in psychiatric settings (Lucksted & Coursey, 1995).
Psychiatric patients experience admissions, including involuntary
admissions, as less coercive when they were implemented with
procedural justice (Lidz, Hoge, Gardner, & Bennett, 1995). More
broadly, in both psychiatric and medical settings, patients place a
premium on negotiation and participatory decision making about
treatment (Cooper-Patrick et al., 1999; Kaplan et al., 1996; Ware,
Tugenberg, & Dickey, 2004). Good relationships require caring,
respectful dialogue about treatment decisions.
In the present study, we developed and validated a measure of
relationship quality in mandated treatment that emphasizes inter-
personal dimensions of both affiliation (bond, caring, and trust)
and control (voice, respect, and fairness). These dimensions and
subcomponents are specified later (see Measures). Our broad
premise is that effective relationships in mandated treatment bear
the hallmarks of the traditional therapeutic alliance but emphasize
an interpersonal form of procedural justice, or a “firm but fair”
relational style (Bonta et al., 2000). A chief departure from the
traditional therapeutic alliance lies in the ongoing processes of role
clarification and open discussion of rules, or what is and is not
negotiable (Trotter, 1999). In effective relationships, this process
unfolds in an authoritative manner that defines and strengthens the
This premise informed a multiple informant design that was
implemented in a specialty probation program. Our draft measure
of relationship quality distilled literature on the therapeutic alli-
ance, mandated treatment, and dual-role relationships (Skeem et
al., 2003). We expected our refined measure of relationship quality
in mandated treatment to be reliable; to predict compliance with
the rules; and to manifest a theoretically coherent pattern of
relationships with ratings of within-session behavior and with
measures of relationships, treatment motivation, and psychological
distress (for hypotheses, see Measures below). We also expected
our measure of dual-role relationship quality to relate more
strongly than a leading measure of the therapeutic alliance to
indices of within-session behavior and relationship satisfaction and
to better predict future rule compliance.
Overview of Studies
We asked pairs of specialty probationers and their officers to
complete our new Dual-Role Relationship Inventory (DRI), along
with measures of theoretically related constructs, shortly after a
regularly scheduled supervision meeting. We audiotaped, tran-
scribed, and then coded transcripts of these supervision meetings
to provide an objective yardstick of the DRI’s ability to capture the
quality of interactions. We also reviewed probationers’ records to
code their compliance with the rules. These data allowed us to
refine the DRI and assess its reliability, validity (e.g., association
with within-session behavior), and predictive utility for rule com-
pliance. We also conducted a secondary study to cross-validate the
structure of the revised DRI. This involved administering the
measure to a larger sample of probationers than we were able to
include in the primary study.
Primary Study
Participant recruitment proceeded in two steps. First, we re-
cruited probation officers within a large and prototypic (see Skeem
et al., 2006) specialty agency in the Southwest by presenting the
study to 11 officers at a monthly staff meeting. Although agency
policies prohibited offering officers an incentive for participating,
7 (63%) officers agreed to do so. Participating officers did not
differ significantly from those who declined to participate in
gender or years of experience. Second, we recruited probationers
on the caseloads of participating officers. All probationers were
diagnosed with an Axis I major mental disorder and were required
to participate in treatment. Study eligibility requirements for pro-
bationers included the following: (a) English speaking, (b) com-
Page 3
petent to provide informed consent, and (c) having met at least
three times with, or having been supervised for at least 2 months
by, the supervising officer. Prospective participants were randomly
selected for recruitment from the current caseloads of participating
officers. As recruitment progressed, probationers were sampled to
match the specialty population in gender, ethnicity, and whether
they had a telephone number (a rough index of financial stability).
Of the 109 participants invited to participate, 12% refused, and 5%
could not be located. The vast majority (83%, n 90) agreed to
Officers (n 7) predominantly were White (100%) women
(71%; men, 29%) with a bachelor’s degree (71%; master’s de-
grees, 29%). Some 14% were of Hispanic ethnicity. Officers were
an average age of 40.8 years (SD 9.9) and had an average of 6.3
(SD 3.1) years’ experience as an officer. The majority (57%)
had prior mental health experience, having worked as a case
manager (75%) or psychologist (25%). Each officer was associated
with 11–14 probationers enrolled in the study.
Of probationers (n 90), the majority were White (64%, Black,
20%, other, 15%) men (61%; women, 39%) with an average age of
37.5 years (SD 9.1). Some 16.9% were of Hispanic ethnicity.
Although most (73.4%) had attained at least a high school degree,
the vast majority (80%) were unemployed, and typically cited
psychiatric disability (68%) as the reason for unemployment. Ac-
cording to their records, probationers’ most common primary
diagnoses were bipolar disorder (34.4%); schizophrenia, schizoaf-
fective, and other psychotic disorders (29.2%); major depression
(24.7%); or other (12.7%) disorders. Of those with an Axis I
mental disorder, 76.7% had one or more co-occurring substance-
related disorders. Most (85.6%) were prescribed psychotropic
medication, and most (77.5%) were also required to complete
substance abuse treatment.
On the basis of their records, probationers had an average of 3.9
(SD 3.8) prior convictions before the index term of probation.
Their most serious charge for the current term of probation was for
a drug (40%), property (24%), person (22%), or minor (13%)
offense. At the time of the study, probationers had spent an
average of 28.1 (SD 30.0; Mdn 24.0) months on probation,
about 9.5 (SD 8.7; Mdn 6.0) months of which was spent with
the officer they rated in the study.
Measures tapped four content domains: the officer–probationer
relationship, officer–probationer interactions, probationers’ inter-
nal state (symptoms and motivation), and probationers’ compli-
ance behavior (probation violations and new arrests). These mea-
sures crossed four sources of information: probationer self-report,
officer report, observer ratings, and record review.
DRI. DRI items were developed by considering the results of
our focus group study (Skeem et al., 2003) and practice guidelines
for treating involuntary clients (Trotter, 1999) in light of the
content of (a) an existing measure for assessing probation officer–
probationer relationship quality (Bonta et al., 2000), and (b) two
existing measures of therapist–client relationship quality, the Cal-
ifornia Psychotherapy Alliance Scales (Marmar, Horowitz, Weiss,
& Marziali, 1986), and the Agnew Relationship Measure (Agnew-
Davies, Stiles, Hardy, Barkham, & Shapiro, 1998). The leading
alliance measure was reserved for use as a validation instrument.
The DRI items were designed to assess two hypothesized do-
mains of relationship quality in mandated treatment. The first is
alliance, which includes bond (acceptance, support, and trust),
partnership (officer efforts to engage the probationer; collaborative
work on problems), and confident commitment (belief and invest-
ment in the helpfulness of the process). Relative to extant scales,
the alliance domain emphasizes the probationers’ trust in the
officer and concern about disclosing information (see Hatcher &
Barends, 1996). The second hypothesized domain is relational
fairness or clarity and voice (clear explanation of limits; freedom
to express opinions), considerate respect (matter of fact application
of rules), and flexible consistency (reasonable accommodation of
rules to individual with consistent enforcement). Notably, the DRI
assesses officers’ use of empathy and warmth to selectively rein-
force prosocial behavior (Bonta et al., 2000; Trotter, 1999).
For each of the two hypothesized domains, 29 questions were
written, for a total of 58 items. Within these domains, items were
positively and negatively worded to minimize the potential influ-
ence of response biases. The Flesch–Kinkaid reading level of the
instrument was grade 4.9.
The DRI has parallel forms for officers, probationers, and ob-
servers (e.g., “My officer considers my views,” “I consider
________’s views,” “The officer considers the probationer’s
views”). In this study, each of the three groups of respondents
indicated how often each item described the target (officer, pro-
bationer, or relationship) on a 7-point, anchored Likert scale that
ranged from 1 (never)to7(always).
Working Alliance Inventory (WAI). The WAI (Horvath &
Greenberg, 1986) is the most frequently used measure of the
therapeutic alliance. The measure consists of 36 items in parallel
form for client, therapist, or observer ratings on a 7-point Likert
scale (never to always). The WAI relates in a theoretically coher-
ent manner with such other variables as treatment outcome (Hor-
vath, 1994). It was adapted for use in this study (with permission)
by replacing the terms therapy, therapist (counselor), and client
with the terms probation, probation officer, and probationer, re-
WAI total scores for probationers (␣⫽.82, M interitem r .12)
and officers (␣⫽.80, M interitem r .10) were used in this study
to assess the convergent validity of the DRI. We expected the WAI
to be moderately associated with the DRI’s alliance domain but
only weakly associated with its relational fairness domain.
Relationship satisfaction. To assess overall relationship sat-
isfaction, we asked probationers and officers to rate on a 5-point
Likert scale, “how satisfied are you in the relationship you have
with (your officer or this probationer)?” Conceptually, relation-
ship satisfaction should be positively associated with relation-
ship quality. Given that the DRI may better capture relationship
quality in mandated treatment, we expected the DRI to be more
strongly associated with general relationship satisfaction than
the WAI.
Page 4
Within-Session Behavior
To assess the nature of officer–probationer interactions, we
arranged 83 meetings between officers and probationers that were
audiotaped, transcribed, and coded by trained observers for inter-
personal process. (Although 90 probationers were interviewed,
technical difficulties rendered seven audiotapes unusable.) The
average length of the audiotaped officer–probationer sessions was
22 min (SD 14.3). Data were prepared for coding by segmenting
transcripts into 50-unit counts on a tape recorder, which corre-
sponds to roughly 4 min of meeting time. Then, each segment was
coded for the presence or absence of six types of officer behavior
and two types of probationer behavior.
The six forms of officer behavior recorded were reflect (state-
ments that reflect content or meaning offered by the probationer),
affirm (statements that complement the probationer’s efforts or
characteristics), support (understanding, supportive, reassuring, or
compassionate comment not captured by reflect or affirm), advise
(giving advice, making a suggestion, or offering a possible action),
direct (giving an order or command), and confront (disagreeing,
contradicting). The two forms of probationer behavior recorded
were change talk (statements that indicate moving forward, in the
direction of compliance with conditions) or resistance (statements
that are inconsistent with or show movement away from compli-
ance). We expected DRI ratings to relate positively to indices of
supportive or positive officer (reflect, affirm, support, advise) and
probationer (change talk) process and negatively to indices of
directive or negative officer (direct, confront) and probationer
(resist) behavior. We also expected the DRI to relate to these
indices more strongly than the WAI.
These behaviors were rated using a coding manual derived for
this study from the Motivational Interviewing Skill Code (Miller,
2000). The manual precisely defined each type of behavior to be
coded and included at least one example of that behavior, drawn
from the study’s audiotapes. For example, direct was defined as
“The officer gives an order or command to the probationer. Or, the
officer has directed the probationer to do something. The language
should be imperative, i.e. must, can’t, should . . . the tone is often
emphatic.” An example of direct was “You need to call the clinic
if you’re going to miss an appointment.”
To ensure adherence to the coding manual, we had raters com-
plete a 2-day workshop, independently rate eight or more training
cases until they reached a good level of agreement with the
criterion (defined as kappa .65), and then meet with the rating
group biweekly to review additional cases to avoid rater’s drift.
During the baseline training, the group met in between each case
to review each rater’s consistency with the criterion ratings and
discuss any discrepancies. Kappa was used to compute interrater
agreement for within-session behavior, using the four most recent
training cases. Generally, kappa values of .75 and greater are
considered to reflect excellent agreement; .60 –.74, good agree-
ment; .40–.59, fair agreement; and .00 –.40, poor agreement (Cic-
chetti & Sparrow, 1981). Using these categorizations, raters found
the team’s average reliability for ratings of within-session behavior
to be excellent (M kappa .75, SD .19).
After rating within-session behavior, raters completed the ob-
server form of the DRI. Raters were trained to reliability on the
DRI using the same process and training cases described earlier.
On the basis of the last four training cases, the team’s average
reliability for DRI ratings was good (weighted kappa .65, SD
.07; Cohen, 1968).
Probationer’s Internal State
Psychological distress. Psychological distress was assessed
using the Brief Symptom Inventory (BSI; Derogatis & Melisara-
tos, 1983), a 53-item self-report inventory in which participants
rate on a 5-point scale the extent to which they have been bothered
(0, not at all to 4, extremely) in the past week by various symp-
toms. The BSI generally manifests a theoretically coherent pattern
of association with the scales of the MMPI (Derogatis & Melisara-
tos, 1983). It includes a measure of global psychological distress,
the general severity index (GSI), and nine subscales designed to
assess individual symptom constellations. However, only four of
the BSI’s subscales demonstrate adequate discriminant validity:
depression, anxiety, somatization, and hostility (for a review, see
Skeem, et al., 2007). In the present study, we used scores on the
full scale (GSI; ␣⫽.96, M interitem r .32) and the depression,
anxiety, and hostility subscales (␣⫽.78-.91, M interitem r
.34 –.61). We expected indices other than hostility to be unrelated
to DRI scores. Ideally, the DRI would work similarly across levels
of distress and negative affect. However, we expected hostility
(which taps anger; Skeem et al., 2006) to relate inversely to DRI
Treatment motivation. Treatment motivation was assessed us-
ing the Situational Motivation Scale (SMS; Guay, Vallerand, &
Blanchard, 2000), a 15-item self-report measure with a 7-point
Likert (1, not at all true to 7, exactly true) scale. The SMS is
designed to assess intrinsic versus extrinsic motivation for a given
activity—in this case, participation in treatment—across four di-
mensions: amotivation (no sense of purpose for treatment partic-
ipation; “There may be good reasons to be in treatment, but
personally I don’t see any”), external regulation (participating in
treatment because of consequences and rewards; “I am in treat-
ment because it is something that I have to do”), intrinsic motiva-
tion (participating in treatment for the inherent pleasure and sat-
isfaction of it; “I am in treatment because it is pleasant”), and
identified regulation (valuing treatment and perceiving it as chosen
by oneself; “I am in treatment because I think it is good for me”).
The SMS is reliable and valid, with scales related as expected to
measures of perceived competence, concentration, and behavioral
intentions (Guay et al., 2000). We expected DRI scores to relate
positively to identified regulation and internal treatment motiva-
tion, given the results of our focus group study (Skeem et al.,
Rule Compliance
We coded recent and future probation violations because the
quality of officer–probationer relationships should relate to rule
compliance. First, probationers’ willingness to comply with the
conditions of probation likely is influenced by the perceived fair-
ness of their enforcement (see MacCoun, 2005) and their rapport
with the officer (Klockars, 1972). Second, negative relationships
can serve as a stressor that compromises probationers’ functioning
and ability to comply with conditions of probation (Skeem et al.,
2003). On the basis of these notions and work suggesting that a
firm but fair (not permissive) approach reduces recidivism (D.
Page 5
Andrews & Kiessling, 1980; Trotter, 1999), we expected the DRI
(which captures relational fairness) to relate more strongly to
future violations and revocation than the WAI (which assesses
only the therapeutic alliance).
Recent violations. Recent violations were defined as those that
occurred during the 2 months preceding the baseline interview.
They were assessed by reviewing probation records to indicate
whether or not there was a recent violation for (a) treatment
noncompliance (psychotropic or psychosocial treatment; base
rate 12.2%), (b) substance use (alcohol or drug possession, use,
or sale; base rate 15.6%), (c) failure to report to the officer (base
rate 12.2%), (d) other technical violation (e.g., failure to pay
fines, pursue work as directed; base rate 43.5%), or (e) physical
violence or commission of a new offense (base rate 2.2%).
Across violation types, nearly half (47.5%) of probationers had
one or more recent violations. Beyond these categorical measures,
the total number of recent violations was recorded, ranging from 0
(53.5%) to 5 (1.1%).
Future violations and probation revocation. To explore the
utility of the DRI in predicting rule compliance, both absolutely
and relative to the WAI, we coded probation records to assess
violations that occurred after the DRI was completed. The average
length of follow-up after baseline interviews was 16.2 months (SD
2.9). At the time of follow-up, half (48.9%) of probationers
were still on specialty mental health probation.
Violations (date of earliest violation, violation type, and number
of violations) and revocation (yes–no) were used as follow-up
variables. The majority of probationers (63.3%) had at least one
violation during the follow-up period (range 1–19, Mdn 1).
The base rates of violations by type were as follows: (a) treatment
noncompliance (33.3%), (b) substance use (47.8%), (c) failure to
report to the officer (23.3%), (d) other technical violation (37.8%),
or (e) physical violence or commission of a new offense (43.3%).
Probation was revoked for nearly one-third of probationers
(32.3%) for a technical violation (15.6%), new offense (3.3%), or
both (13.3%).
Research assistants (RAs) recruited probationers by mail, tele-
phone, home visits, and probation office visits. RAs arranged to
meet probationers expressing interest in the study at the probation
office at their next regularly scheduled meeting with their officer.
Prior to the officer–probationer meeting, the RA obtained in-
formed consent from probationers, had probationers complete the
symptom measure (the BSI), and arranged a tape recorder in the
officer’s office. Following the officer–probationer meeting, the
RA met with the probationer to complete the study materials,
provided officers with the study materials to complete within 24
hr, and coded probationers’ records. For both probationers and
officers, the order of study materials was counterbalanced to avoid
order effects. Probationer participants were paid $50. Audiotaped
officer–probationer meetings were transcribed and rated by trained
Secondary Study
In the secondary study, we administered a revised version of the
DRI to a large sample of probationers to cross-validate the struc-
ture of the measure identified in the first study. Participants were
322 probationers with mental disorder drawn from a specialty
mental health agency in a large Southern city and a traditional
agency in a large Western city. Eligibility criteria were similar to
those applied in the primary study. Of probationers invited to
participate, most (78%) agreed to do so. As a group, participants
were an average of 36.7 years old (SD 10.8), with the majority
being male (58%; female, 42%) and Black (50%; White, 28%;
other, 12%). Most (74.5%) were also required to participate in
psychiatric or substance abuse treatment as a condition of proba-
tion. According to their records, probationers’ most common pri-
mary diagnoses were bipolar disorder (44.0%); schizophrenia,
schizoaffective, and other psychotic disorders (30.5%); major de-
pression (20.4%); or other (5.1%) disorders. Probationers’ most
serious charge for their lifetime history of arrest was for a person
(53%), drug (24%), property (19%), or minor (4%) offense. At the
time of the study, probationers had spent an average of 8.6 (SD
14.4) months on probation, about 4.6 (SD 6.2) months of which
was spent with the officer they rated in the study.
Probationers completed a revised version of the DRI (DRI–R
described below) that included 30 items. Given indications during
the secondary study that probationers with cognitive impairments
had difficulty completing negatively worded items, these items
were revised in the middle of this secondary study to be positively
worded. Because (a) the measures used in Study 1 and Study 2
differ and (b) the measure used in Study 2 was revised well into
recruitment, these are noisy data that provide a relatively stringent
cross-validation of the structure of the DRI–R.
RAs recruited eligible probationers by mail, telephone, home
visits, and probation office visits. RAs arranged to meet proba-
tioners expressing interest in the study at probation offices, public
places, or probationers’ residences. After securing informed con-
sent, the RA completed an interview with the probationer that
included administration of the DRI–R. Participants were paid $40.
The data from both studies were analyzed in four phases. First,
data from the primary study were prepared for analysis by “cen-
tering around officers” to reduce problems of data dependence.
Second, the internal structure of the DRI was analyzed using
primary study data to refine the measure by deleting items with
questionable distributions and relationships with the overall scale.
Third, the structure of the DRI–R was explored via confirmatory
factor analysis, using data from the secondary study. Fourth, the
reliability, validity, and predictive utility of the DRI–R was ana-
lyzed using primary study data, with emphasis on the measure’s
relation to within-session behavior, treatment adherence, and rule
compliance. In this section, each phase is described.
Preparing the Primary Study Data
Data included 90 officer ratings, 90 probationer ratings, and 83
observer ratings. For cases missing less than 10% of DRI items,
Page 6
missing values were replaced with average values for that case.
Because seven officers were associated with 90 probationers, there
was a potential for biased officer ratings (in which particular
officers rated their relationships as more positive or negative than
the officer group) or affected probationer or observer ratings (in
which probationers or observers nested within an officer rated
relationships as more positive or negative than their group). Mul-
tivariate analyses of variance and interclass correlation analyses
(see Kenny & La Voie, 1985) indicated that officers, but not
probationers or observers nested within officers, had systematic
rating biases. To reduce problems of data dependence, we centered
officers’ DRI scores to remove systematic response biases. That is,
an individual officer’s average DRI score across his or her cases
was deducted from his or her rating of each case, and then the
officer group average rating was added to each item score (see
Hatcher, Barends, Hansell, & Gutfreund, 1995). Analyses con-
ducted with and without officer-centered scores produced the same
pattern of results.
Refining the DRI
After the data were prepared for analyses, the DRI was refined
on the basis of analyses of its internal structure. First, distributions
of responses to DRI items were examined within rater type (offi-
cer, probationer, or observer) to identify items with poor distribu-
tions. Items with poor distributions were defined as those with (a)
very high or low average scores and low standard deviations, (b)
extreme splits at the median response (i.e., 8% in either direc-
tion), or (c) visibly skewed or kurtotic distributions. Second, the
degree of association among the items was computed to identify
items that were weakly associated with the remaining item pool.
Specifically, we computed squared multiple correlations between
each item and the remaining item pool, bivariate correlations
among the items, and exploratory factor analysis. Given the lim-
ited sample size within rater types (N 83–90), factor analyses
were repeated within and across observer types to identify a stable
solution. A three-factor oblique solution was most stable within the
three observer types and across observer type combinations
(officer–probationer and officer–probationer– observer groups).
Items that failed to contribute to this solution were considered for
deletion or revision.
This solution was coherent, but not neatly consistent with the
hypothesized DRI domains of alliance and fairness. Instead, the
first and largest factor was marked by bond (alliance) and clarity–
voice (fairness) items. The second factor consisted of items that
tapped the extent to which the probationer and officer trusted one
another from the bond (alliance) items. The third factor tapped the
officers’ toughness and punitiveness with the probationer (e.g.,
disciplinary orientation; expectations of independence) on the ba-
sis of items from a variety of designed scales. On the basis of these
results, the factors were provisionally labeled Caring–Fairness,
Trust, and Toughness.
After integrating the results of these item correlation and item
distribution analyses and revisiting theories of dual-role relation-
ships, 28 items were deleted from the DRI. The deleted items
chiefly were from the hypothesized alliance scales that were de-
signed to assess partnership and confident commitment (see Mea-
sures above). These domains may be less relevant to relationship
quality in mandated treatment than in traditional psychotherapy.
The results were also used to revise the wording of other items to
better target DRI–R domains. These revised items are included in
the current, 30-item DRI (DRI–R).
Assessing the Structure of the DRI–R
Data on 322 probationers from the secondary study were used to
cross-validate the exploratory factor structure of the DRI–R ob-
served in the primary study. Amos 5.0.1 was used to test the
correlated three-factor model depicted in Figure 1. Notably, this
structure is mathematically equivalent to a hierarchical three-factor
structure in which a superordinate relationship quality factor over-
arches the three factors. We depict the correlated version to show
the degree of association among each of the three factors, which is
attributable to the superordinate factor. Comparative fit index
(CFI) values at or above .90 and root-mean-square error of ap-
proximation (RMSEA) values at or below .10 were used to define
adequate fit (Browne & Cudeck, 1993; Byrne, 1994). Applying
those criteria, the three-factor model manifested an acceptable fit
to our sample’s data,
(402, N 322) 1481.11, p .001,
RMSEA .09, CFI .90. The basic three-factor model fit the
data better than (a) a two-factor model in which Trust items were
combined with those of Caring–Fairness to create one factor, and
Toughness remained a separate factor,
(433, N 322)
1654.21, p .001, RMSEA .09, CFI .88, and (b) a one-factor
(405, N 322) 1762.28, p .001, RMSEA .10,
CFI .87.
Assessing Psychometric Properties of the DRI–R
Given these results, in the remainder of this article, we describe
the psychometric properties of the DRI–R, with the original DRI
item wording and with the items grouped into three DRI–R scales
that reflect the sum of Caring–Fairness, Trust, and Toughness
items shown in Figure 1. Total scores were the sum of the items,
reversing those of the Toughness scale. We examine DRI–R scales
individually because use of a multidimensional test as a single
variable can obscure important relationships with criterion vari-
ables (Smith, Fischer, & Fister, 2003).
Reliability and Validity
With respect to reliability, DRI–R scales and total scores had
excellent internal consistency (␣⫽.96, .90, .87, and .95 for
Caring–Fairness, Toughness, Trust, and Totals, respectively) and
moderate interitem correlations (Mr .59, .67, .56, and .59,
respectively). To assess cross-informant reliability and convergent
validity, we analyzed the pattern of correlations between and
within rater types for total scores on the DRI–R, total scores on the
WAI, and global ratings of relationship satisfaction. As shown in
Table 1, there was poor cross-informant agreement across rater
types for all measures. For example, probationers’ ratings on the
DRI–R (r .06) and the WAI (r .17) were not significantly
associated with their officers’ ratings on the same measures. As
shown later, these findings are consistent with literature on the
therapeutic alliance. At the subscale level, the pattern of cross-
informant correlations (not shown) makes theoretical sense. For
example, observer and officer, but not probationer, ratings of
Page 7
probationers’ Trust are correlated (r .29, p .01). Similarly,
observer and probationer, but not officer, ratings of officers’
Toughness are correlated (r .28, p .05).
Construct Validity
The within-rater data provided in Table 1 provide evidence of
convergent validity for the DRI–R. First, within rater types, the
DRI–R and WAI were moderately associated. For example, offic-
ers’ WAI ratings were moderately (r .42) correlated with their
DRI–R ratings, suggesting that the DRI–R taps something that is
related to, but not the same as, the therapeutic alliance. Second,
within rater types, the DRI–R is more strongly associated with
global ratings of satisfaction than is the WAI. For example, offic-
ers’ satisfaction ratings are not significantly associated with their
WAI scores (r .18) but weakly to moderately associated with
their DRI–R scores (r .37). This suggests that the DRI–R taps
qualities that are more relevant to respondents’ appraisals than the
These data are consistent with analyses of the relation between
ratings of within-session behavior and scores on the WAI and
DRI–R. No significant relationships were found between within-
session behavior and probationers’ or officers’ total scores on the
WAI. In contrast, within-session behavior bore a theoretically
coherent pattern of relationships with probationers,’ officers,’ and
observers’ DRI–R total and scale scores, as shown in Table 2. For
example, all informants’ total DRI–R scores were moderately
strongly inversely associated with officers’ confrontation of pro-
bationers in session and moderately inversely associated with
probationers’ resistance of officers in session. Observers’ DRI–R
scores were moderately positively related to affirming and sup-
portive within-session behavior by the officers. A closer look at the
pattern of association between ratings of within-session behavior
and scales of the DRI–R provides some evidence for the scales’
validity. For example, within-session ratings of resistance by pro-
bationers were positively associated with probationers’ DRI–R
ratings of Toughness (r .41), inversely associated with Trust
(r ⫽⫺.33), and weakly and nonsignificantly associated with
Caring–Fairness (r ⫽⫺.18).
Next, we assessed the relation between the DRI–R and treatment
motivation. The results are shown in Table 3. We expected DRI–R
scores to be significantly associated with probationers’ amotiva-
tion, which was the case for some subscale scores (e.g., probation-
ers’ and observers’ ratings of Toughness), but not total scores.
Similarly, although we expected DRI–R total scores to relate
positively to intrinsic motivation, there was merely a trend at the
scale level (for observers) and significant findings for two sub-
.83 a s about m as a personXc re e
.78 explains what I am supposed to do & why… X
.84 X tries very hard to do the right thing by me
.83 When I’m having trouble, X talks with me…
.74 If I break the rules, X calmly explains what…
.81 X is enthusiastic and optimistic with me
.87 X encourages me to work with him or her
.86 X really considers my situation when…
.85 X seems devoted to helping me overcome…
.85 X is warm and friendly with me
.87 X treats me fairly
.90 X really cares about my concerns
.78 X praises me for the good things I do
.76 If I’m going in a bad direction, X will talk…
.86 I know that X truly wants to help me
.85 X considers my views
.83 X gives me enough of a chance to say…
.86 X takes enough time to understand me
.87 X takes my needs into account
.79 X expects me to do all the work alone…
.77 X makes unreasonable demands of me
.87 I feel that X is looking to punish me
.78 X puts me down when I’ve done something…
.76 X talks down to me
.78 I feel safe enough to be open and honest…
.77 I feel free to discuss the things that worry…
.78 X trusts me to be honest with him or her
.75 X knows that he or she can trust me
.87 X is someone I trust
.85 X shows me respect in absolutely all his/her…
Figure 1. Factor structure of the revised Dual-Role Relationship Inventory (DRI–R).
Page 8
scales (e.g., probationers’ and observers’ ratings of Trust and
Toughness, respectively).
The chief measure of divergent validity in the study was the BSI
measure of psychological distress. Ideally, DRI–R scores would be
relatively independent of a probationers’ symptomatology. As
hypothesized, we found that probationers’, officers’, and observ-
ers’ total DRI–R scores were not significantly associated with the
BSI indices of general distress, depression, anxiety, or somatiza-
tion. Probationer DRI–R total scores, but not those of officers or
observers, were significantly inversely associated with the BSI
Hostility scale (r ⫽⫺.21, p .05), as predicted. Aside from the
BSI Hostility scale, there was only 1 (of 36) significant relation
between the DRI–R subscales and the BSI subscales: Probationers’
DRI–R Toughness scores were weakly associated with their BSI
ratings of anxiety (r .22, p .05).
Predictive Utility for Rule Compliance
First, the association of the DRI–R with recent probation vio-
lations was assessed. As shown in Table 4, officers’ DRI–R total
scores and both officers’ (Trust) and probationers’ (Toughness)
subscale scores were significantly associated with the number of
recent violations. The pattern of association differed by violation
Second, we assessed the utility of the DRI–R in predicting
whether, and how quickly, probationers violated one or more
conditions of probation. Because probationers’ time at risk varied
as a function of when their interview was completed (see Measures
above), Cox proportional hazards survival analyses were com-
pleted to determine whether the four DRI–R scales predicted
probationers’ time to their first violation (or lack thereof). The
scales were entered in a forward stepping algorithm, with the
likelihood ratio as the criterion for entry and removal. The results
indicated that time to violation was predicted by one or more
DRI–R subscales completed by probationers,
(1, N 90)
5.59, p .01; officers,
(1, N 90) 6.02, p .01; and
(1, N 83) 7.71, p .01. For probationers,
officers, and observers, only the Toughness scale, exp() 1.28,
p .01, Caring–Fairness scale, exp() 0.26, p .01, and Trust
scale, exp() 0.68, p .01, respectively, were significantly
predictive. These estimated odds ratios mean, for example, that for
every 1-point increase in a probationers’ DRI–R Toughness score,
the odds of violation increased by 28%. A repetition of these
analyses with DRI–R Total scores as the only predictor indicates
that officers’,
(1, N 90) 3.77, p .05, but not probationers’
or observers’ global scores predicted time to violation. WAI scores
did not predict time to violation.
Third, parallel survival analyses were completed to assess the
utility of the DRI–R in predicting whether, and how quickly,
probation was revoked for rule noncompliance that was deemed
serious. As with violations, the subscale-based analyses indicated
that time to revocation was predicted by one or more DRI–R
subscales completed by probationers,
(1, N 90) 12.51, p
.01; officers,
(1, N 90) 7.26, p .01; and observers,
N 83) 10.14, p .01. For observers, the Caring–Fairness
scale, exp() 0.32, p .01, and Trust scale, exp() 0.40, p
.01, significantly predicted time to revocation. For probationers
and officers, only the Toughness scale, exp() 1.94, p .01,
and Caring–Fairness scale, exp() 0.12, p .01, respectively,
were significantly predictive. These estimated odds ratios mean,
Table 1
Cross-Informant Agreement on Relationship Quality and Within-
Informant Association of Relationship Measures
Probationer Officer
WAI DRI-R Satisfaction WAI DRI-R Satisfaction
DRI-R .54
Satisfaction .27
WAI .13 .11 .13
DRI-R .17 .06 .19 .42
Satisfaction .01 .03 .11 .18 .37
DRI-R .16 .07 .10 .13 .26
Note. WAI Working Alliance Inventory; DRI-R revised Dual-Role
Relationship Inventory.
p .05.
p .01.
Table 2
Relation Between Informants’ DRI-R Scores and Ratings of Within-Session Behavior
Probationer scores Officer scores Observer DRI-R
Total C-F/Trust/Tough Total C-F/Trust/Tough Total C-F/Trust/Tough
Officer behavior
Reflect .05 .06/.04/.01 .08 .11/.00/.08 .23
Affirm .02 .00/.01/.15 .03 .02/.08/.03 .43
Support .13 .14/.13/.03 .09 .04/.16/.01 .37
Advise .02 .01/.05/.11 .01 .00/.02/.02 .21 .23
Direct .04 .07/.01/.05 .28
Confront .22
Probationer behavior
Change talk .07 .02/.05/.26
.10 .01/.12/.18 .11 .10/.18/.02
Resist .27
Note. DRI-R revised Dual-Role Relationship Inventory; C-F Caring-Fairness subscale; Trust Trust subscale; Tough Toughness subscale.
p .05.
p .01.
p .001.
Page 9
for example, that for every 1-point increase in a probationers’
DRI–R Toughness score, the odds of revocation increased by 94%.
A repetition of these analyses with DRI–R total scores as the only
predictor indicates that officers’,
(1, N 84) 4.86, p .05,
but not probationers’ or observers’ global scores predicted time to
revocation. Neither officers’ nor probationers’ WAI scores pre-
dicted time to revocation. In summary, the DRI–R is linked with
both past and future rule compliance, ranging from recent treat-
ment noncompliance to future revocation.
In this study, we refined and validated a measure of relationship
quality in mandated treatment. The DRI-R was developed in a
context in which care and control are equally emphasized, with
specialty mental health probation officers and their supervisees
with mental disorder. The three primary findings of this study have
implications for understanding, assessing, and evaluating the effect
of dual-role relationships. First, the results indicate that the quality
of dual-role relationships is not adequately captured by traditional
conceptualizations of the therapeutic alliance. In the context of
mandated treatment, effective relationships involve not only car-
ing, but also fairness, trust, and an authoritative (not authoritarian)
style. Here, caring becomes blended with fairness. Second, the
study indicates that the DRI–R assesses these key domains, is
internally consistent, and manifests a theoretically coherent pattern
of associations with ratings of within-session behavior and mea-
sures of the therapeutic alliance, relationship satisfaction, symp-
toms, treatment motivation, and future rule compliance. Moreover,
the DRI–R was more strongly related to within-session behavior
and relationship satisfaction and more strongly predictive of rule
compliance than the leading measure of the therapeutic alliance.
Third, the utility of the DRI–R in predicting probation violations
and new arrests suggests that the quality of dual-role relationships
affects important outcomes for those with co-occurring disorders.
Beyond technical interventions, the process of care and supervi-
sion is crucial. In this section, we describe each of these main
Understanding Dual-Role Relationships
The results of this study indicate that the quality of dual-role
relationships is not adequately captured by traditional conceptual-
izations of the therapeutic alliance. The nature of relationship
quality differs in traditional and dual-role contexts, perhaps be-
cause the latter context involves primary interpersonal dimensions
of both affiliation and control (see Benjamin, 1996). There are
three key differences between the therapeutic relationship and the
dual-role relationship. First, although caring (bond, attachment) is
key to both types of relationships, a collaborative willingness to
invest in the therapy process seems less central to dual-role rela-
tionships. Of DRI items designed to tap the alliance, those that
tapped bond (acceptance, support, trust, openness) were retained,
whereas those that tapped partnership (working together toward
shared goals) and confident commitment (belief in one another’s
abilities) were deleted because they did not cohere with the rest of
the scale. This is consistent with Angell and Mahoney’s (2007)
finding that the themes discussed by case managers in focus
groups “pertain to aspects of the relationship bond, but did not
focus specifically upon goal and task consensus, as would be
Table 3
Relation Between Informants’ DRI-R Scores and Ratings of Within-Session Behavior
Situational Motivational
Scale subscale
Probationer DRI-R scores Officer DRI-R scores Observer DRI-R scores
Total C-F/Trust/Tough Total C-F/Trust/Tough Total C-F/Trust/Tough
Amotivation .17 .12/.12/.31
.06 .14/.05/.14 .14 .11/.12/.22
External Regulation .15 .15/.13/.14 .21
.02 .01/.07/.02
Intrinsic Motivation .11 .17/.21
.03 .12/.08/.00 .19 .18/.12/.22
Identified Regulation .05 .01/.07/.15 .10 .09/.03/.09 .23
Note. DRI-R revised Dual-Role Relationship Inventory; C-F Caring-Fairness subscale; Trust Trust subscale; Tough Toughness subscale.
p .05.
p .01.
Table 4
Relation Between Informants’ DRI-R Scores and Recent Recorded Rule Compliance
Recent violations
Probationer DRI-R scores Officer DRI-R scores Observer DRI-R scores
Total C-F/Trust/Tough Total C-F/Trust/Tough Total C-F/Trust/Tough
Number .15 .09/.10/.34
.08 .04/.13/.19
Treatment noncompliance .02 .04/.04/.07 .17 .08/.22
.14 .10/.17/.18
Substance abuse .13 .09/.11/.23
.18 .15/.16/.25
Failure to report .26
.16 .09/.19/.08 .07 .10/.03/.00
Other technical .08 .04/.12/.19 .23
.04 .06/.04/.07
Violence/new offense .19
.19 .19/.07/.12 .11 .11/.05/.13
Note. DRI-R Dual-Role Relationship Inventory; C-F Caring-Fairness subscale; Trust Trust subscale; Tough Toughness subscale.
p .05.
p .001.
Page 10
predicted by the literature on the working alliance” (p. 31). In
dual-role relationships, the power imbalance may obviate the need
for clients to agree with treatment and supervision decisions.
The second key difference is that, in dual-role relationships,
caring becomes blended with fairness. In the DRI–R, caring and
fairness are integrated into one scale that is associated with trust.
This finding may be viewed as a relational variant of the finding
that citizens’ trust in the law, or compliance with decisions made
by authority figures, depends on their perception of both proce-
dural justice and benevolent or caring motives (Tyler & Huo,
2002). In contrast with our expectations that they would be rela-
tively distinct, the aspects of the alliance that are relevant to
dual-role relationships are integral components of relational jus-
Third, the style of implementing control appears to be a key
component of dual-role relationships, but not the therapeutic alli-
ance. Beyond Caring–Fairness and Trust, a darker side of dual-role
relationships emerged in this study that we labeled Toughness. As
shown in Figure 1, the items of this scale tap a punitive orientation
and expectations of independence and compliance. This scale
seems to bode poorly for relationships: It was associated with
officer confrontation within sessions, probationer mistrust, treat-
ment amotivation, and future rule noncompliance. Angell and
Mahoney (2007) found that limit setting in dual-role case man-
agement relationships introduced a parent– child-like dynamic into
the relationship. Along these lines, the Toughness scale seems to
capture an authoritarian supervisory style. Research on parenting
suggests that authoritarian parents are highly demanding and di-
rective but (unlike authoritative parents) are not responsive to their
children’s needs. “They are obedience- and status-oriented, and
expect their orders to be obeyed without explanation” (Baumrind,
1991, p. 52). The Toughness scale emphasizes an indifference to
probationers’ views and feelings, expectation of compliance, and
punitiveness when expectations are not met. This darker side of
dual-role relationships is consistent with Angell and Mahoney’s
(2007) qualitative finding that conflict and struggle over issues of
power and control are an important negative aspect of case man-
agement relationships. The nature of the relationship is different in
dual-role contexts than in strictly therapeutic ones: In mandated
treatment, there is an emphasis on caring, fairness, trust, and an
authoritative (not authoritarian) style.
Assessing Dual-Role Relationships
The DRI–R, which captures these dimensions, is the first vali-
dated measure of dual-role relationship quality developed to date.
Relative to the most widely used measure of the therapeutic
alliance (the WAI), the DRI–R relates more strongly to ratings of
within-session behavior and officers’ and probationers’ satisfac-
tion with their relationships. This suggests that the DRI–R captures
facets that are more relevant to the nature and quality of relation-
ships in mandated treatment than are traditional measures of the
With respect to reliability, the DRI–R is internally consistent at
both total score and subscale levels. Although there was low
convergence between probationers’ and officers’ (but not officers’
and observers’) ratings on the DRI–R, there was also low conver-
gence between their ratings of the WAI. Indeed, this finding of
poor cross-rater agreement is consistent with findings on the
therapeutic alliance in psychotherapy (e.g., Tichenor & Hill,
1989). Across studies in psychotherapy, each perspective appears
similarly predictive of treatment outcome (Martin et al., 2000),
with the client perspective having a slight advantage (Horvath &
Symonds, 1991). Similarly, in the present study, officers’ and
probationers’ total DRI–R ratings were both predictive of rule
noncompliance (although for different reasons). Within each per-
spective (probationer, officer, or observer), DRI–R ratings corre-
lated with other variables (e.g., within-session behavior) in a
theoretically coherent manner. Despite different views of the re-
lationship, it seems that each view possesses some validity.
The DRI–R also related to such variables as the treatment
alliance, relationship satisfaction, within-session behavior, treat-
ment motivation, and future rule compliance in a theoretically
coherent manner, both at the total score and scale level. For
example, officers’ confrontation of probationers within session
was strongly positively linked with observers’ rating of Tough-
ness, strongly negatively linked with their ratings of Trust, and
moderately negatively related to their ratings of Caring–Fairness.
The DRI–R was developed and validated in a mandated treat-
ment context that involves prototypically dual roles. In the rela-
tionship between specialty mental health officers and probationers
with co-occurring disorders, casework and supervision are both
paramount. The extent to which the DRI–R will generalize to
capture the quality of relationships between clinicians and invol-
untary clients is unclear. Nevertheless, two points bode positively
for the measure’s generalizability. First, there are a number of
parallels between work on specialty mental health probation
(Roskes & Feldman, 1999; Skeem et al., 2003; Solomon et al.,
2002) and others’ work on intensive case management (Angell &
Mahoney, 2007; Neale & Rosenheck, 2000). For example, both
contexts involve social control, limit setting, and potential conflict
over issues of power and compliance. As noted by McCabe and
Priebe (2004), “statutory responsibilities for care and the require-
ment to monitor patients in the community . . . mean that many
‘therapeutic relationships’ are initiated and maintained not by the
patient but by the mental health professional, a feature of assertive
outreach models of care and all forms of ‘compulsory treatment’”
(p. 124). Second, the most widely used measure of the therapeutic
alliance (the WAI) has been extended with some success well
beyond the traditional psychotherapy relationship (see Introduc-
tion, above). Recently, the WAI was used to capture the parent–
therapist alliance, which predicted positive outcomes in parent
management training (Kazdin, Whitley, & Marciano, 2006). It is
possible that the DRI–R may generalize to mandated treatment
contexts beyond probation (e.g., case managers or therapists with
involuntary clients; counselors with substance abuse clients; men-
tal health court judges with defendants). Nevertheless, because
particular types of relationships in mandated treatment may differ
in their tasks, settings, formality, and weighting of clinical versus
supervisory roles, the generalizability of the DRI–R must be di-
rectly investigated in future research.
Evaluating the Impact of Dual-Role Relationships
In keeping with our past work (Skeem et al., 2003), we found
that the quality of the dual-role relationship predicted rule com-
pliance, that is, probation violations, probation revocation, and
new arrests. This was not the case for the therapeutic alliance. The
Page 11
mechanism for this association is unclear. It is possible that dual-
role relationship quality links directly with rule compliance,
through procedural justice (MacCoun, 2005), the skills of the
synthetic officer (Klockars, 1972), or both. First, Tom Tyler’s
(1990; Tyler & Huo, 2002) work on procedural justice indicates
that citizens’ willingness to comply with decisions made by legal
authorities is linked with perceptions that these decisions are based
on fair procedures and benevolent motives. Such perceptions are
created by treating people with dignity, respect, and caring. These
are aspects of positive dual-role relationships tapped by the
Second, the skills of the synthetic officer may be at work.
According to Carl Klockars’ (1972) ethnography, the synthetic
type of officer obtains compliance by satisfying both therapeutic
and social control roles. In contrast, the law enforcer and time
server are unshakably rule-enforcing officers, whereas the thera-
peutic agent errs in the direction of unstructured support and
permissiveness. Only the synthetic approach brings the power of
both the relationship and the mandate to bear on rule compliance.
In this study, synthetic relationships (indexed by DRI–R total
scores) predicted rule compliance, whereas therapeutic agent rela-
tionships (indexed by WAI scores) were unrelated to future rule
compliance. These findings are consistent with D. Andrews and
Kiessling’s (1980) finding that a firm but fair approach (e.g., one
that involves both active listening and directive supervision) is
most effective in reducing recidivism risk for general probationers.
Authoritarian and permissive approaches are less effective than
authoritative ones.
Synthetic relationships and relational justice describe a link
between dual-role relationship quality and rule compliance that is
direct. However, there is probably an indirect link, as well,
wherein the effect of relationship quality on rule compliance is
mediated by mental state. Negative relationships can compromise
probationers’ mental state, functioning, and ability to comply with
the conditions of probation. As noted by one probationer with
mental disorder:
My mental condition is something of a severe emotional turbulence
. . . and anything that causes me an additional bit of unease or
anything, you know, additionally bad in my life, contributes to the
strain of a situation that is already teetering on the brink of suicide. So
. . . it seems like it would make sense for my probation officer . . . to
be very decent in his treatment of me. . .. (Skeem et al., 2003, pp.
The bulk of the literature on social support and social under-
mining indicates that relative to positive social exchanges, nega-
tive ones more strongly predict deterioration in well-being and
quality of life (Baumeister, Bratslavsky, Finkenhauser, & Vohs,
2001; Rook, 1998; Rosenfield & Wenzel, 1997). In turn, compro-
mised mental state and functioning may limit probationers’ ability
to comply with the conditions of probation (e.g., getting to re-
quired appointments; abstaining from substance use; Skeem et al.,
2003). Indeed, there is evidence that negative treatment alliances
and clinicians’ limit setting elevate patient’s risk of becoming
involved in violence (Beauford, McNiel, & Binder, 1997; see also
Fagan-Pryor et al., 2003; Ilkiw-Lavalle & Grenyer, 2003). Thus,
negative dual-role relationships may compromise probationers’
mental state and ability to comply with rules.
Whatever the mechanism, it is clear that the quality of dual-role
relationships predicts rule compliance. In the context of specialty
mental health probation, rule compliance relates directly to
whether probationers with co-occurring disorders are sent to
prison. This population is rapidly growing and is at high risk
deepening involvement in the criminal justice system: Compared
with their relatively healthy counterparts, probationers with mental
disorder are twice as likely to violate probation and return to
incarceration (Dauphinot, 1997). Training specialty officers and
others with dual-role relationships to establish caring, fair, trusting,
and nonauthoritarian relationships may improve outcomes for
these probationers. To improve outcomes for high-risk populations
in mandated treatment, we must include the process of supervision
and treatment in our definition of evidence-based practice.
Abram, K. M., & Teplin, L. A. (1991). Co-occurring disorders among
mentally ill jail detainees: Implications for public policy. American
Psychologist, 46, 1036 –1045.
Agnew-Davies, R., Stiles, W. B., Hardy, G. E., Barkham, M., & Shapiro,
D. A. (1998). Alliance structure assessed by the Agnew Relationship
Measure (ARM). British Journal of Clinical Psychology, 37, 155–172.
Alverson, H., Alverson, M., & Drake, R. E. (2000). An ethnographic study
of the longitudinal course of substance abuse among people with severe
mental illness. Community Mental Health Journal, 36, 340–345.
Andrews, D., & Kiessling, J. (1980). Program structure and effective
correctional practice: A summary of CaVic research. In R. Ross & P.
Gendreau (Eds.), Effective correctional treatment (pp. 439 463). To-
ronto, Ontario, Canada: Butterworths.
Andrews, D. A., Zinger, I., Hoge, R. D., & Bonta, J. (1996). Does
correctional treatment work? A clinically relevant and psychologically
informed meta-analysis. In D. F. Greenberg (Ed.), Criminal careers
(Vol. 2, pp. 437– 472). Brookfield, VT: Dartmouth.
Angell, B., & Mahoney, C. (2007). Reconceptualizing the case manage-
ment relationship in intensive treatment: A study of staff perceptions and
experiences. Administration and Policy in Mental Health and Mental
Health Services Research, 34, 172–188.
Angell, B., Mahoney, C. A., & Martinez, N. I. (2006). Promoting adher-
ence in assertive community treatment. Social Service Review, 80,
Asay, T. P., & Lambert, M. J. (1999). The empirical case for the common
factors in therapy: Quantitative findings. In M. A. Hubble, B. L. Duncan,
& S. D. Miller (Eds.), The heart and soul of change: What works in
therapy (pp. 23–55). Washington, DC: American Psychological Asso-
Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001).
Bad is stronger than good. Review of General Psychology, 5, 323–370.
Baumrind, D. (1991). The influence of parenting style on adolescent
competence and substance use. Journal of Early Adolescence, 11, 56
Beauford, J. E., McNiel, D. E., & Binder, R. L. (1997). Utility of the initial
therapeutic alliance in evaluating psychiatric patients’ risk of violence.
American Journal of Psychiatry, 154, 1272–1276.
Benjamin, L. S. (1996). A clinician-friendly version of the Interpersonal
Circumplex: Structural analysis of social behavior (SASB). Journal of
Personality Assessment, 66, 248 –266.
Bonta, J., Hanson, R. K., Sedo, B., Coles, R., Nascimento, G., & Rugge, T.
(2000). Effective case management in Manitoba probation. Unpublished
project description.
Brown, P. D., & O’Leary, K. D. (2000). Therapeutic alliance: Predicting
continuance and success in group treatment for spouse abuse. Journal of
Consulting and Clinical Psychology, 68, 340–345.
Page 12
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model
fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation
models (pp. 136 –162). Newbury Park, CA: Sage.
Bureau of Justice Statistics. (2006). Correctional facts at a glance. Re-
trieved August 5, 2006, from
Byrne, B. M. (1994). Structural equation modeling with Eqs and Eqs/
Windows: Basic concepts, applications, and programming. Thousand
Oaks, CA: Sage.
Cicchetti, D., & Sparrow, S. (1981). Developing criteria for establishing
interrater reliability of specific items: Applications to assessment of
adaptive behavior. American Journal of Mental Deficiency, 86, 127–
Cohen, J. (1968). Weighted kappa: Nominal scale agreement provision for
scaled disagreement or partial credit. Psychological Bulletin, 70, 213–
Connors, G. J., Carroll, K. M., DiClemente, C. C., Longabaugh, R., &
Donovan, D. M. (1997). The therapeutic alliance and its relationship to
alcoholism treatment participation and outcome. Journal of Consulting
and Clinical Psychology, 65, 588–598.
Cooper-Patrick, L., Gallo, J. J., Gonzales, J. J., Vu, H. T., Powe, N. R.,
Nelson, C., & Ford, D. E. (1999). Race, gender, and partnership in the
patient–physician relationship. Journal of the American Medical Asso-
ciation, 282, 583–589.
Cruz, M., & Pincus, H. A. (2002). Research on the influence that commu-
nication in psychiatric encounters has on treatment. Psychiatric Services,
53, 1253–1265.
Dauphinot, L. L. (1997). The efficacy of community correctional supervi-
sion for offenders with severe mental illness (Doctoral dissertation,
University of Texas at Austin, 1997). Dissertation Abstracts Interna-
tional, 57, 5912.
Derogatis, L. R., & Melisaratos, N. (1983). The Brief Symptom Inventory:
An introductory report. Psychological Medicine, 13, 595–605.
Ditton, P. (1999). Mental health and treatment of inmates and probation-
ers. Washington, DC: Bureau of Justice Statistics.
Drake, R. E., Wallach, M. A., Alverson, H. S., & Mueser, K. T. (2002).
Psychosocial aspects of substance abuse by clients with severe mental
illness. Journal of Nervous and Mental Disease, 190, 100–106.
Fagan-Pryor, E. C., Haber, L. C., Dunlap, D., Nall, J. L., Stanley, G., &
Wolpert, R. (2003). Patients’ view of causes of aggression by patients
and effective interventions. Psychiatric Services, 54, 549–553.
Frank, A. F., & Gunderson, J. G. (1990). The role of the therapeutic
alliance in the treatment of schizophrenia: Relationship to course and
outcome. Archives of General Psychiatry, 47, 228–236.
Guay, F., Vallerand, R. J., & Blanchard, C. (2000). On the assessment of
situational intrinsic and extrinsic motivation: The Situational Motivation
Scale. Motivation and Emotion, 24, 175–213.
Hall, J. A., Horgan, T. G., Stein, T. S., & Roter, D. L. (2002). Liking in the
physician–patient relationship. Patient Education and Counseling, 48,
68 –77.
Hartwell, S. (2004). Triple stigma: Persons with mental illness and sub-
stance abuse problems in the criminal justice system. Criminal Justice
Policy Review, 15(1), 84 –99.
Hatcher, R., & Barends, A. (1996). Patients’ view of the alliance in
psychotherapy: Exploratory factor analysis of three alliance measures.
Journal of Consulting and Clinical Psychology, 64, 1326–1336.
Hatcher, R. L., Barends, A., Hansell, J., & Gutfreund, M. J. (1995).
Patients’ and therapists’ shared and unique views of the therapeutic
alliance: An investigation using confirmatory factor analysis in a nested
design. Journal of Consulting and Clinical Psychology, 63, 636 643.
Hellerstein, D. J., Rosenthal, R. N., & Miner, C. R. (1995). A prospective
study of integrated outpatient treatment for substance-abusing schizo-
phrenic patients. American Journal on Addictions, 4, 33–42.
Henry, W. P., & Strupp, H. H. (1994). The therapeutic alliance as inter-
personal process. In A. O. Horvath & L. S. Greenberg (Eds.), The
working alliance: Theory, research, and practice (pp. 51– 84). Oxford,
England: Wiley.
Horvath, A. O. (1994). Empirical validation of Bordin’s pantheoretical
model of the alliance: The Working Alliance Inventory perspective. In
A. O. Horvath & L. S. Greenberg’s (Eds.), The working alliance:
Theory, research, and practice (pp. 109–128). New York: Wiley.
Horvath, A., & Greenberg, L. (1986). The development of the Working
Alliance Inventory. In L. Greenberg & W. Pinsof (Eds.), The psycho-
therapeutic process: A research handbook (pp. 529 –556). New York:
Guilford Press.
Horvath, A., & Luborsky, L. (1993). The role of the therapeutic alliance in
psychotherapy. Journal of Consulting and Clinical Psychology, 61,
Horvath, A., & Symonds, B. (1991). The role of the therapeutic alliance in
psychotherapy. Journal of Consulting and Clinical Psychology, 38,
139 –149.
Howgego, I. M., Yellowlees, P., Owen, C., Meldrum, L., & Dark, F.
(2003). The therapeutic alliance: The key to effective patient outcome?
A descriptive review of the evidence in community mental health case
management. Australian and New Zealand Journal of Psychiatry, 37,
169 –183.
Ilkiw-Lavalle, O., & Grenyer, B. F. (2003). Differences between patient
and staff perceptions of aggression in mental health units. Psychiatric
Services, 54, 389 –393.
Kaplan, S. H., Greenfield, S., & Ware, J. E. (1989). Assessing the effects
of physician–patient interactions on the outcomes of chronic disease.
Medical Care, 27, S110 –S127.
Kaplan, S. H., Sullivan, L. M., Spetter, D., Dukes, K. A., Khan, A., &
Greenfield, S. (1996). Gender patterns of physician–patient communi-
cation. In M. M. Falik, & K. S. Collins (Eds.), Women’s health: The
Commonwealth Fund Survey (pp. 76 –95). Baltimore, MD: Johns Hop-
kins University Press.
Kazdin, A. E., Whitley, M., & Marciano, P. L. (2006). Child–therapist and
parent–therapist alliance and therapeutic change in the treatment of
children referred for oppositional, aggressive, and antisocial behavior.
Journal of Child Psychology and Psychiatry, 47, 436 445.
Kenny, D. A., & La Voie, L. (1985). Separating individual and group
effects. Journal of Personality and Social Psychology, 48, 339–348.
Klockars, C. (1972). A theory of probation supervision. The Journal of
Criminal Law, Criminology, and Police Science, 63, 550–557.
Krupnick, J., Sotsky, S., Simmens, S., Moyer, J., Elkin, I., Watkins, J., &
Pilkonis, P. (1996). The role of the therapeutic alliance in psychotherapy
and pharmacotherapy outcome: Findings in the National Institute of
Mental Health Treatment of Depression Collaborative Research Pro-
gram. Journal of Consulting and Clinical Psychology, 64, 532–539.
Lidz, C. W., Hoge, S. K., Gardner, W., & Bennett, N. S. (1995). Perceived
coercion in mental hospital admission: Pressures and process. Archives
of General Psychiatry, 52, 1034–1039.
Luborsky, L., Rosenthal, R., Diguer, L., Andrusyna, T. P., Berman, J. S.,
Levitt, J. L., et al. (2002). The dodo bird verdict is alive and well—
Mostly. Clinical Psychology: Science and Practice, 9, 2–12.
Lucksted, A., & Coursey, R. D. (1995). Consumer perceptions of pressure
and force in psychiatric treatments. Psychiatric Services, 46, 146–152.
MacCoun, R. J. (2005). Voice, control, and belonging: The double-edged
sword of procedural fairness. Annual Review of Law and Social Science,
1, 171–201.
Marmar, C. R., Horowitz, M. J., Weiss, D. S., & Marziali, E. (1986). The
development of the Therapeutic Alliance Rating System. In L. S. Green-
berg, & W. M. Pinsof (Eds.), The psychotherapeutic process: A research
handbook (pp. 367–390). New York: Guilford Press.
Martin, D., Garske, J., & Davis, M. (2000). Relation of the therapeutic
alliance with outcome and other variables: A meta-analytic review.
Journal of Consulting and Clinical Psychology, 68, 438 450.
Page 13
McCabe, R., & Priebe, S. (2004). The therapeutic relationship in the
treatment of severe mental illness: A review of methods and findings.
International Journal of Social Psychiatry, 50, 115–128.
Miller, W. R. (2000). Motivational Interviewing Skill Code. Retrieved
August 5, 2006, from
Monahan, J., Steadman, H. J., Robbins, P. C., Appelbaum, P., Banks, S.,
Grisso, T., et al. (2005). An actuarial model of violence risk assessment
for persons with mental disorders. Psychiatric Services, 56, 810 815.
Neale, M. S., & Rosenheck, R. A. (2000). Therapeutic limit setting in an
assertive community treatment program. Psychiatric Services, 51, 499
Penn, D. L., Mueser, K. T., Tarrier, N., Gloege, A., Cather, C., Serrano, D.,
& Otto, M. W. (2004). Supportive therapy for schizophrenia: Possible
mechanisms and implications for adjunctive psychosocial treatments.
Schizophrenia Bulletin, 30, 101–112.
Rook, K. S. (1998). Investigating the positive and negative sides of
personal relationships: Through a lens darkly? In B. H. Spitzberg &
W. R. Cupach (Eds.), The dark side of close relationships (pp. 369
393). Mahwah, NJ: Erlbaum.
Rosenfield, S., & Wenzel, S. (1997). Social networks and chronic mental
illness: A test of four perspectives. Social Problems, 44, 200–216.
Roskes, E., & Feldman, R. (1999). A collaborative community-based
treatment program for offenders with mental illness. Psychiatric Ser-
vices, 50, 1614 –1619.
Roter, D. L., & Hall, J. A. (1992). Doctors talking with patients, patients
talking with doctors. Westport, CT: Auburn House.
Skeem, J. L., Emke-Francis, P., & Eno Louden, J. (2006). Probation,
mental health, and mandated treatment: A national survey. Criminal
Justice and Behavior, 33, 158–184.
Skeem, J. L., Encandela, J., & Eno Louden, J. (2003). Perspectives on
probation and mandated mental health treatment in specialized and
traditional probation departments. Behavioral Sciences & the Law, 21,
429 458.
Skeem, J., Schubert, C., Odgers, C., Mulvey, E., Gardner, W., & Lidz, C.
(2007). Psychiatric symptoms and community violence among high-risk
patients: A test of the relationship at the weekly level. Journal of
Consulting and Clinical Psychology, 74, 967–979.
Smith, G. T., Fischer, S., & Fister, S. M. (2003). Incremental validity
principles of test construction. Psychological Assessment, 15, 467– 477.
Solomon, P., Draine, J., & Marcus, S. C. (2002). Predicting incarceration
of clients of a psychiatric probation and parole service. Psychiatric
Services, 53, 50 –56.
Taft, C. T., Murphy, C. M., King, D. W., Musser, P. H., & DeDeyn, J. M.
(2003). Process and treatment adherence factors in group cognitive–
behavioral therapy for partner violence men. Journal of Consulting and
Clinical Psychology, 71, 812– 820.
Tichenor, V., & Hill, C. (1989). A comparison of sic measures of the
working alliance. Psychotherapy, 26, 195–199.
Trotter, C. (1999). Working with involuntary clients: A guide to practice.
Thousand Oaks, CA: Sage.
Tyler, T. R. (1990). Why people obey the law. New Haven, CT: Yale
University Press.
Tyler, T. R., & Huo, Y. J. (2002). Trust in the law: Encouraging public
cooperation with the police and courts. New York: Russell Sage Foun-
U.S. Probation and Pretrial Services. (2001). Court and community [“Fact
Sheet” on probation]. Washington, DC: Federal Judicial Center.
Ware, N. C., Tugenberg, T., & Dickey, B. (2004). Practitioner relationships
and quality of care for low-income persons with serious mental illness.
Psychiatric Services, 55, 555–559.
Wolfe, B., & Goldfried, M. (1988). Research on psychotherapy integration:
Recommendations and conclusions from an NIMH workshop. Journal of
Consulting and Clinical Psychology, 56, 448 451.
Received August 7, 2006
Revision received June 1, 2007
Accepted June 5, 2007
Page 14
    • "and were averaged to create a composite score for perceived support, voice, fairness, and respect (Mean = 5.5, SD = 1.6). Next, a 5-item questionnaire adapted from the Dual Role Relationship Inventory (Skeem, Louden, Polaschek, & Camp, 2007) was used to assess parents' perceptions of the probation officer's helpfulness toward the youth. These five items include—the probation officer tries to make sure my child doesn't get lost between the cracks of the systems; treats my child with respect; ignores my child's needs; goes the extra mile to help my child; and genuinely wants to help my child. "
    [Show abstract] [Hide abstract] ABSTRACT: In the past several years, there has been a growing movement toward family-driven initiatives in many child-serving agencies, including the juvenile justice system. These initiatives underscore the importance of parental involvement in successful rehabilitation of at-risk and offending youth and highlight the unique role of parents to influence and inspire their child's behavior. Despite a growing consensus on the importance of parental involvement in juvenile justice processes, little empirical research has explored the nature of parental involvement in the juvenile justice system. This study examined parents' (n = 87) perceptions of relationship quality and interaction with probation officers, parenting strategies, and how these factors related to youth's compliance on probation. Findings revealed that parents generally had positive relationships with probation officers characterized as supportive, fair, respectful, and helpful toward youth. Most parents also employed practices such as use of reminders and encouragement to promote youth's compliance on probation. Parents' perceptions of probation officers' helpfulness toward youth were associated with decreased used of parenting practices that encourage probation compliance. However, parents' perceptions of supportive, respectful, and fair relationships with probation officers were associated with increased use of parenting practices that promote probation compliance. Supportive, fair, and respectful relationships with probation officers were also linked to fewer counts of technical violations of probation, but not new delinquent offenses, among offending youth. Implications for research, practice, and policy around the potential of collaborative relationships between parents and probation officers in facilitating successful probation outcomes are discussed.
    No preview · Article · Apr 2016 · Children and Youth Services Review
  • Source
    • "Law (2016 Skeem & Manchak, 2008). Further, synthetic officers are thought to be better able to establish high-quality relationships with their supervisees, which is a predictor of lower recidivism among offenders (Kennealy, Skeem, Manchak, & Eno Louden, 2012; Skeem, Eno Louden, Polaschek, & Camp, 2007; Skeem & Manchak, 2008). These findings suggest that officers' attitudes toward their work and the offenders they supervise can have important consequences for offenders (see also Gendreau & Ross, 1987; Kennealy et al., 2012). "
    [Show abstract] [Hide abstract] ABSTRACT: This research examined how probation officers use risk information about offenders, and how its use is affected by what aspects of their role they emphasize. Officers (N = 152) were invited to complete surveys before and after a risk assessment tool training (46.0-65.8% participation rate). Surveys assessed estimates of reoffense and officers' likely supervision approach given a probationer's risk level. Officers tended to overestimate the likelihood of medium- and high-risk offenders to reoffend. As risk level rose, officers tended to increase the number of meetings and referrals. Officers' role emphases were related to how they perceived low-risk offenders' likelihood to reoffend, but not for offenders at other risk levels. There was relative consistency in officers' role emphases, supervision decisions, and responses to violations, so that differences in practice did not appear to be systematic. Most officers' risk perceptions were more realistic after training. Copyright © 2016 John Wiley & Sons, Ltd.
    Full-text · Article · Mar 2016 · Behavioral Sciences & the Law
  • Source
    • "For instance, little is known regarding how to instil protective factors in our clients. Moreover, we presume a focus on strengths will enhance the working relationship, an important component of client success (Kennealy, Skeem, Eno Louden, & Manchak, 2012; Skeem, Eno Louden, Polaschek, & Camp, 2007), but this has not been empirically confirmed. Informing risk assessment with the systematic use of standardised protocols should advance our understanding regarding key risk and protective constructs. "
    [Show abstract] [Hide abstract] ABSTRACT: This paper reviews the literature regarding the identification and measurement of risk factors considered imminent precipitants of subsequent criminal conduct (i.e. dynamic risk factors). This paper also frames these risk factors against the so-called protective factors that are presumed to mitigate risk. Commonality among recent dynamic risk and protective measures reflects general agreement regarding viable candidate variables. Empirical studies suggest such factors yield incremental predictive validity and should inform case-formulaic understanding of criminal conduct and pathways to desistance, although this is not common practice. As well, definitional and measurement considerations are not well advanced and speak to the need for further conceptual clarity.
    Preview · Article · Nov 2015 · Psychology Crime and Law
Show more