Received: 12 December 2017 Revised: 22 March 2018 Accepted: 23 March 2018

DOI: 10.1002/cpp.2197

RESEARCH ARTICLE

A therapist version of the Alliance Negotiation Scale

Jennifer M. Doran1,2 | Juan Martín Gómez‐Penedo3 | Jeremy D. Safran4 | Andrés Roussos3

1
VA Connecticut Healthcare System, CT, USA
2
Abstract
Department of Psychiatry, Yale School of
Medicine, CT, USA The aim of the current study was to design and evaluate a therapist version of the
3
CONICET, Universidad de Buenos Aires, Alliance Negotiation Scale (ANS). The ANS was created in order to operationalize
Buenos Aires, Argentina
4
the construct of dyadic negotiation in psychotherapy and to augment existing concep-
The New School for Social Research, New
York, NY, USA tualizations of the working alliance. The ANS has existed only as a client self‐report
Correspondence form since its inception and has demonstrated promise as a psychotherapy process
measure. This research intended to develop a complementary therapist self‐report
Jennifer M. Doran, Ph.D., VA Connecticut
Healthcare System, CT, USA version of the measure. The scale creation process is discussed in detail, and the
Email: [email protected]
results of a preliminary psychometric investigation are reported. The ANS‐Therapist
version (ANS‐T) was developed using a sample of therapists (n = 114) through a prin-
cipal components analysis procedure. The ANS‐T contains 9 unidimensional items and
was moderately correlated with therapist‐reported working alliance (r = .468). The
results of the study support the composition of the ANS‐T and provide initial support
for the reliability and validity of the measure.

KEY W ORDS

Alliance Negotiation Scale, psychotherapy process measure, therapist self‐report, working alliance

1 | I N T RO D U CT I O N Their foundational work emerged in the context of the relational ther-

apy model (e.g., Mitchell & Aron, 1999) and has focused more on neg-
The Alliance Negotiation Scale (ANS; Doran, Safran, Waizmann, ative interpersonal processes in therapy. The importance of tension,
Bolger, & Muran, 2012) was created in 2012 in an effort to disagreements, and upset feelings between a patient and therapist
operationalize the theoretical construct of dyadic negotiation in psy- began to be seen as critically important aspects of the therapeutic pro-
chotherapy (Safran & Kraus, 2014; Safran & Muran, 2000, 2006) and cess (Safran, Crocker, McMain, & Murray, 1990). Such negative ther-
in order to augment existing conceptualizations of the working alli- apy events are known as ruptures in the therapeutic relationship
ance. Bordin is credited with developing the modern definition of (Safran, 1993), and a body of literature has grown supporting the
the working alliance, framed as a collaborative stance between a importance of ruptures and their resolution (Safran & Muran, 2000),
patient and therapist emerging from agreement on the tasks and goals their relationship to treatment outcome (Aguirre McLaughlin, Keller,
of therapy, as well as the quality of their relational bond (Bordin, Feeny, Youngstrom, & Zoellner, 2014; Safran, Muran, & Eubanks‐Car-
1979). Although the working alliance has become a foundational con- ter, 2011), and their role in the change process (Stiles et al., 2004;
cept in the practice and study of psychotherapy and has decades of lit- Strauss et al., 2006).
erature to support it, historical conceptualizations are not without The ANS was developed in this context and the theoretical
their limitations or critiques. The working alliance and its measurement argument to augment the current conceptualization of the working
are generally positively valenced in nature, with the hallmark feature alliance to include the constructive dyadic process of negotiation
of existing measures described as “collaboration and consensus” over the more traditional focus on agreement and compliance (for
(Cushman & Gilford, 2000; Horvath, Re, Fluckiger, & Symonds, a more detailed review of this and aforementioned issues, see
2011). Safran and Muran (2006) have argued that the more traditional Doran, 2016; Doran, Safran, & Muran, 2016; Safran & Muran,
concept has become somewhat outdated and have pushed for the 2000, 2006). Alliance negotiation represents what patients and ther-
definition to evolve to include more modern theories and concepts. apists do when ruptures or other negative processes occur in

Clin Psychol Psychother. 2018;25:745–753. wileyonlinelibrary.com/journal/cpp © 2018 John Wiley & Sons, Ltd. 745
746 DORAN J. M. ET AL.

therapy sessions and reflect if the quality of the therapeutic relation-

ship facilitates or hinders these issues being addressed and resolved. Key Practitioner Message:
Theoretically, a high level of dyadic negotiation is expected to be
related to positive treatment outcome, whereas problems in negoti- • Alliance negotiation is emerging as an important topic of

ation are expected to negatively impact the therapy process and study in psychotherapy research

outcome. Constructive alliance negotiation is conducive to identify- • The Alliance Negotiation Scale (ANS) and ANS‐
ing and working through ruptures and other negative therapy pro- Therapist version may be important tools for tracking
cesses, paving the way for positive therapeutic and interpersonal alliance negotiation in psychotherapy
change. Research is beginning to support these ideas (Doran, Safran, • The ANS and ANS‐T may provide clinicians with
& Muran, 2017). important information about the quality of the
Although there are numerous existing patient and therapist mea- therapeutic relationship and help identify potential
sures of psychotherapy process (e.g., the Individual Therapy Process problems in negotiation
Questionnaire; Mander et al., 2015) and the working alliance specifi-
• The ANS and ANS‐T can be used by clinicians to track
cally (see Bachelor, 2013, for an overview of the three most common
alliance negotiation throughout treatment and may
measures of the alliance), the ANS is viewed as offering a unique con-
serve as an important bidirectional feedback tool
tribution in its explicit focus on negative therapy process related to
• The ANS and ANS‐T can help clients and therapists
patient/therapist feelings, behaviour, and perceptions of the other
identify their ability and willingness to work through
and their dyad during treatment. Although some existing alliance mea-
negative feelings and difficult therapy process when it
sures may contain a few items that assess the presence of negative
occurs
thoughts or feelings, the ANS is unique in its focus on what the client
and therapist actually do with these feelings when they occur. The
ANS moves beyond assessing the presence of disagreement, for
example, to examining how the client and therapist respond to and
address disagreement when it occurs. The ANS is also the first mea- Results of a preliminary study on the ANS‐A were comparable with
sure to explicitly operationalize and assess the construct of alliance that of the original ANS, demonstrating relative cultural equivalence
negotiation. and potential utility of the construct cross‐culturally (Waizmann
The ANS was modelled in structure and form after the Working et al., 2015). Both English and Spanish versions of the ANS have
Alliance Inventory (WAI; Horvath & Greenberg, 1986, 1989), one of continued to demonstrate psychometric integrity through investiga-
the most commonly used measures of the working alliance, with the tions of their reliability as well as content, criterion, and construct
intention to augment the measurement of the alliance construct by validity (Doran et al., 2012; Doran et al., 2016; Doran et al., 2017;
including an explicit focus on negative therapy process and the pres- Gomez Penedo, 2017; Roussos, Gomez Penedo, Doran, Olivera, &
ence and resolution of ruptures in therapy (Doran et al., 2016). Horowitz, 2017).
Although the original WAI has been transformed into several ver- Studies using the ANS have demonstrated relationships between
sions, including shorter forms as well as therapist self‐report and negotiation and other process and outcome variables of interest,
observer‐based coding versions, the ANS was initially created only including both client and therapist factors. Client self‐reported inter-
as a client self‐report version of the measure (Doran et al., 2012). It personal problems, such as nonassertiveness and social avoidance,
was determined that the reliability, validity, and utility of the con- as well as the presence of a personality disorder, were associated
struct of negotiation, and the ANS as an independent measure, with lower levels of negotiation in the therapeutic relationship (Doran
should be examined before investing resources into the creation of et al., 2016; Doran et al., 2017). Client perceptions of therapist empa-
parallel versions. thy were positively associated with negotiation (Doran et al., 2016).
The ANS is a 12‐item client self‐report measure consisting of two Lower levels of negotiation have been associated with more client‐
factors: Comfort with Negative Feelings (Factor 1) and Flexible and identified ruptures in session, as well as with client behaviours aimed
Negotiable Stance (Factor 2). ANS items assess both client feelings at avoiding tension in the therapeutic relationship. In contrast, higher
and behaviour (such as “I am comfortable expressing disappointment levels of negotiation have been found to be associated with the res-
in my therapist when it arises”), client perceptions of the relational olution of ruptures and greater session impact (Doran et al., 2017),
dyad (“I feel that I can disagree with my therapist without harming as well as increased client satisfaction with treatment (Doran et al.,
our relationship”), and client perceptions of the therapist's behaviour 2016). The ANS has also proven useful in helping to explain some
(“My therapist is able to admit when he/she is wrong about something variance in treatment outcome, with higher negotiation scores corre-
we disagree on”). Items are rated on a 7‐point Likert scale, with lated with decreased symptom levels on measures of general psychi-
responses ranging from Never to Always, and Factor 2 items are atric distress and interpersonal functioning. Furthermore, lower ANS
negatively valenced and reverse scored. Appendix A includes a copy scores have demonstrated a predictive relationship with premature
of the measure. termination from treatment (Doran et al., 2017). A study on the
Since its inception, the ANS has been translated into Spanish and ANS‐A found that hostile‐dominant and hostile‐submissive interper-
adapted for use in an Argentinean population (ANS‐A; Waizmann sonal problems were related to lower negotiation scores overall
et al., 2015), with several other cultural adaptations in progress. (Gomez Penedo, 2017). The ANS‐A has also demonstrated a
DORAN J. M. ET AL. 747

relationship with treatment outcome, with higher early negotiation 2 | METHOD

scores predicting improvement in interpersonal problems of low
agency (social avoidance, nonassertiveness, and being overly exploit- Given the popularity of the ANS in both American and Argentinean
able), as well as predicting change on a global outcome measure cultures, the creators of the respective versions of the scale decided
(Roussos et al., 2017). to collaborate on the development of a therapist version of the mea-
Negotiation, as measured by the ANS, has been found to be sure. Traditionally, assessment measures are created in one language
highly correlated with both the working alliance, as measured by the or culture and then tested in another culture for equivalence and util-
WAI, and the real relationship, as measured by the Real Relationship ity or adapted/translated as needed. This approach to assessment
Inventory (Gelso et al., 2005), indicating substantial convergence often results in very Western‐centric tools and places the country of
across constructs and measures (Doran et al., 2012; Doran et al., origin (typically the United States) in a position of privilege and as
2016). The relationship between the ANS and WAI has been found the point of reference for subsequent versions. As the concept of
to vary over the course of treatment, with more distinction occurring negotiation has been found to be useful cross‐culturally, and it is likely
in the early and middle phases of treatment. Also, less overlap has that the ANS‐T will be used in multiple cultural contexts, it was deter-
emerged between the ANS and WAI when comparing “low” versus mined that a collaborative cross‐cultural effort to create a therapist
“high” negotiation cases, indicating that the ANS may capture some version would be preferable. Members of the research teams involved
unique relational difficulties (Doran et al., 2017). in the validation of both the ANS and ANS‐A worked together in order
Differences have not emerged in levels of negotiation across cli- to simultaneously (rather than sequentially) create and develop parallel
ent symptom levels or modalities of treatment, though negotiation versions of the ANS‐T. This paper presents the final English/American
scores have been found to be higher for dyads who had been work- version of the scale; an identical Argentinean version of the measure
ing together for longer periods of time and who met more regularly was also created as part of this process and will be presented in a sep-
for treatment (Doran et al., 2016). Relationships between negotia- arate paper in a Spanish‐language journal (Gomez Penedo, Doran, &
tion and demographic characteristics have varied across studies. Roussos, 2017).
One study found that negotiation scores were slightly higher for
female clients and identified a positive linear relationship between
negotiation and client age. In the same study, clients endorsing a
2.1 | Procedure
racial/ethnic minority identity reported lower negotiation scores The present study was designed to mirror the original ANS develop-
overall, and negotiation was lower in culturally “mismatched” ther- ment study (for a detailed review of this process, see Doran et al.,
apy dyads, for example, a minority client with a White therapist 2012). The first step was to create an item pool for potential items
(Doran et al., 2016). that would comprise the ANS‐Therapist version (ANS‐T). The item
As in any interpersonal occurrence, the therapeutic relationship creation process for the client version of the ANS was a long and iter-
and the working alliance are dyadic constructs. Understanding the ative process that involved collaboration with a core research team of
thoughts, feelings, and perceptions of both client and therapist is clinician‐researchers. Content validity checks also occurred through
needed in order to fully understand the psychotherapy process and review by senior members of the research team and an external expert
any associated relational construct. It is therefore critical to investigate panel of psychotherapy researchers who reviewed and rated all poten-
the presence and perception of negotiation from both the client and tial items as part of the original ANS study (additional information
therapist in determining the impact of negotiation on treatment about the item construction process can be found in Doran et al.,
outcome. Research using previous working alliance measures has 2012). Building on the theoretical groundwork that was part of this
demonstrated a lack of convergence in perceptions of the alliance original item creation process, existing ANS client‐centred items were
across reporters (Hersoug, Hoglend, Monsen, & Havik, 2001; Tichenor directly translated into a parallel item from the perspective of the ther-
& Hill, 1989), with differential impact on the process–outcome rela- apist. For example, in the client version of the ANS Item 1 is “I feel that
tionship as well (Horvath & Symonds, 1991; Piper, Azim, Joyce, & I can disagree with my therapist without harming our relationship.”
McCallum, 1991). The directly translated item for the ANS‐T would therefore be “I feel
The absence of a therapist version of the ANS has limited existing that I can disagree with my patient without harming our relationship.”
research efforts, providing psychotherapy process data only from the Although reasonable translations were able to be made for all original
perspective of the patient. Given this, a therapist version of the ANS ANS items, it was acknowledged that the most direct translation might
is greatly needed in order to fully understand the dyadic process of not provide the clearest or most psychometrically sound way of
alliance negotiation in psychotherapy, as well as to continue to vali- assessing the concept of interest. Therefore, at least one alternate
date the construct of negotiation itself and more clearly determine wording was created for each item in an effort to reframe the item
its relationship with treatment outcome. The aim of the present study wording or increase clarity of the concept. For Item 1, additional
was to create a psychometrically sound therapist version of the ANS. wordings that were tested included “My patient and I are able to con-
It was hypothesized that a therapist version of the ANS would closely structively work through disagreements in therapy” as well as “My
parallel the existing client version in terms of factor structure and con- patient seems to feel that he/she can disagree with me without
tent. Based on research using the client versions of both scales (Doran harming our relationship.” As in the original ANS, items in the initial
et al., 2017), a moderate‐to‐large correlation between the ANS and item pool were framed both to capture the perspective of the individ-
WAI was also expected. ual completing the self‐report (the client in the original version; the
748 DORAN J. M. ET AL.

therapist for the ANS‐T), as well as to assess their perception of the endorsed a sexual minority status (14.1%). Therapist age ranged from
“other” in the dyad. As another example, Item 2 in the original ANS 23 to 75 (M = 44.1, SD = 14.6). Therapists primarily identified as psy-
is “My therapist encourages me to express any concerns I have with chologists (60.5%) or psychology trainees (27.2%). Several respon-
our progress.” For the ANS‐T, parallel items for testing included “I dents also identified as being a psychiatrist, a mental health
encourage my patient to express any concerns he/she has with our counsellor, a social worker, or a psychoanalyst (12.3%). The majority
progress” (direct translation), as well as “I am comfortable hearing my of participants held a doctoral degree (69.3%) and worked in either a
patient's negative feelings about me or our work” (alternate phras- clinical practice (57.0%), hospital or medical centre (28.1%), or other
ing/wording). Although both tested items are similar in nature, one mental health clinic (14.9%). Years of clinical experience ranged from
assesses the therapist's feelings while the other their actual behaviour. 1 to 47 (M = 15.2, SD = 11.5).
This was done purposefully, as the item pool in the original ANS (and The sample was close to evenly split between being assigned to
the actual final version of the measure itself) contains both types of answer the survey questions while thinking about a “difficult”
items. The goal for the ANS‐T was to test items that captured (a) ther- (49.1%) versus an “easy” (50.9%) client. Therapists reported that more
apist feelings, (b) therapist behaviour, and (c) therapist perceptions of than half of their clients were female (57.0%). The majority of clients
the patient/dyad. was also described as White (76.3%), heterosexual (82.5%), and youn-
Participants for the current study were recruited through psychol- ger than their therapists (64.9%). Client diagnoses included depressive
ogy graduate student university departments, listserves for psycholo- disorders (50.9%), anxiety disorders (6.8%), a trauma or stressor‐
gists, and psychologist social networking sites using a snowball related disorder (24.6%), or other clinical syndrome such as an eating
sampling procedure. An initial invitation to participate was sent via disorder or adjustment disorder (24.6%). A subset of the sample was
email or as a discussion post and requested participation from clini- diagnosed with a personality disorder (21.1%).
cians. Participants who were interested in the study were invited to
click on a link to learn more, which directed them to a data collection
website (Survey Monkey) where they were provided with informed 3 | RESULTS
consent and given the option to participate. Informed consent was
obtained from all participants prior to their participation. The survey All data were analysed using the SPSS (version 23) statistical program.
consisted of 60 questions and first assessed therapist demographic All data were checked for errors and screened for outliers prior to run-
and professional information, followed by de‐identified client demo- ning any analyses. Data were also checked to ensure that no statistical
graphic information. This information was collected through the use assumptions were violated that would prevent the use of parametric
of multiple‐choice questions with the option to include a unique, tests in the data analysis (Kendall & Stuart, 1958). Any case with more
open‐ended response where applicable. Clinicians also provided infor- than 5% of missing data was excluded from subsequent analyses. No
mation about their theoretical orientation, modality of treatment, fre- items were excluded from the pool of potential ANS‐T items on the
quency of sessions, duration of treatment, and their client's presenting basis of violating a statistical criterion.
problems and diagnoses.
In order to encourage a range of responses and assess the quality
of the therapeutic relationship in both positive and more challenging
3.1 | Principal components analysis
dyads, participants were randomly assigned to answer subsequent The Kaiser–Meyer–Olkin test of sampling adequacy for factor analytic
questions while thinking about a current client who they perceive as techniques exceeded the recommended threshold (Kline, 1994), pro-
either easy or difficult to work with. Participants then responded to ducing a score of .801. An exploratory principal components analysis
each of the 26 potential ANS‐T items as well as the 12 items from (PCA) was performed on the data in order to identify any underlying
the short version of Working Alliance Inventory–Therapist version dimensions in the scale items and to reduce the item pool down to a
(WAI‐T; Tracey & Kokotovic, 1989) to provide preliminary data on simplified measure. A direct translation of the client version of the
convergent and discriminant validity. Items from the ANS‐T pool and ANS was first tested before examining the differential functioning of
WAI‐T were presented together and randomly interspersed in an directly translated versus alternate items. The final scale was created
effort to control for order effects. Finally, clinicians were asked how through careful analysis of both theoretical and statistical criterion,
much they liked their client overall and how they felt about their work with the goal of retaining the most psychometrically sound items
together given a Likert‐style scale. across both the English‐ and Spanish‐language versions of the scale.
The PCAs were run simultaneously in order to find the best fit for
both sets of data. From a statistical standpoint, established criteria
2.2 | Participants were utilized to qualify an item for retention (Bryant & Yarnold,
All participants (N = 114) were mental health providers who were cur- 1995; Clark & Watson, 1995; Tabachnick & Fidell, 2001). Retained
rently seeing patients under their own licence or as a trainee under the items were required to have an eigenvalue above 1, factor loadings
supervision of a licensed professional. Participants came from 15 above .4, and a minimum difference of .4 on an item that loaded on
countries (n = 20 respondents were from outside of the United States) multiple factors.
and 18 U.S. states. The majority of respondents was female (70.2%), The final solution was obtained through running a PCA with a
heterosexual (80.7%), and White (93%). A minority of respondents one‐factor solution and without rotation. Table 1 presents each item
identified as Asian (4.4%) or Hispanic/Latino (2.6%), as well as with its corresponding factor loading. As with all factor analytic
DORAN J. M. ET AL. 749

TABLE 1 ANS‐T items and factor loadings

ANS‐T item Factor loadings

I believe my patient feels comfortable expressing frustration in me when it arises .78
I believe my patient feels comfortable expressing disappointment in me when it arises .77
My patient and I are able to constructively work through and resolve tension or ruptures in our relationship .76
I encourage my patient to express any anger he/she feels towards me .73
I am comfortable hearing my patient's negative feelings about me or our work .72
I regularly “check in” with my patient to see if he/she feels that the way we are working together is correct .61
I am able to admit to my patient when I am wrong about something we disagree on .59
I feel that I can disagree with my patient without harming our relationship .54
I believe my patient feels like he/she has a say regarding what we do in therapy .46

Note. ANS‐T = Alliance Negotiation Scale–Therapist version.

procedures, multiple solutions were tested as part of the data analysis p = .01. These small correlations can also be taken as evidence of both
process in order to identify the best fit. Utilizing a PCA versus principal convergent and discriminant validity. Although there is some relation-
axis factoring would not have changed the composition of the scale, as ship between the constructs, alliance negotiation also appears to be
both methods yielded highly similar results. Many different versions of fairly distinct from liking the client or overall feelings about the work.
the ANS‐T were also tested and subject to scrutiny, replacing directly In contrast, correlations on these variables with the WAI‐T were
translated items with alternates in a sequential fashion in order to higher, at r = .64, p < .001, and r = .70, p < .001, respectively, indicating
determine the optimal composition of the scale. As the original ANS a much larger degree of overlap and conflation between the con-
includes two orthogonal subscales, and the present item pool was structs. These correlational data are presented in Table 2.
closely modelled after the original items, a forced two‐factor solution
was also tested. Contrary to expectations, this solution was not a good
fit of the data, and the absence of an underlying factor structure in the
3.4 | Treatment characteristics
ANS‐T was confirmed by graphical representations on both the scree Treatment modality varied across therapists, with over half of the sam-
plot and the component plot in rotated space (not shown). The princi- ple identifying as integrative or eclectic (56.1%). A substantial number
ples of psychometric theory (Gregory, 2004) were closely adhered to of therapists endorsed the use of psychodynamic psychotherapy/psy-
in reviewing and evaluating alternate versions of the ANS‐T through- choanalysis (57.9%) and/or cognitive behaviour therapy or another
out the scale construction process. cognitive–behavioural treatment (52.6%). Several also selected human-
istic/existential therapy (19.3%). Given that the majority of therapists
endorsed more than one modality and very few selected only one,
3.2 | The ANS‐T analysing differences on ANS‐T scores between individual types of
The final version of the ANS‐T includes nine unidimensional items. Six treatment was not appropriate in this sample. Most therapy sessions
of these items were direct translations of the original ANS wording, occurred at a frequency of once a week or more (80.7%), and just under
and three of these items included alternate versions of an original item half of the dyads under study had been working together for 1 year or
(ANS Items 2, 7, and 12). Three items were also dropped from the longer (45.6%). No significant differences on ANS‐T scores emerged
measure due to inadequate performance on either the original or when comparing treatment frequency or duration using one‐way anal-
alternate versions of the item (ANS Items 8, 10, and 11). A final ther- yses of variance and independent t tests (all ps > .05).
apist version of the ANS, the ANS‐T, is presented in Appendix B.

3.5 | Demographic and diagnostic characteristics

3.3 | Psychometric properties
Independent t tests, bivariate Pearson correlations, and one‐way anal-
Cronbach's alpha coefficient was adequate for the final version of the yses of variance were run to analyse demographic and diagnostic
ANS‐T, α = .843, exceeding the recommended threshold of .80 (Clark
TABLE 2 Correlations between the ANS‐T and other study variables
& Watson, 1995; Gregory, 2004), and 44.9% of the variance in scores
was explained by the final version of the ANS‐T. A bivariate Pearson ANS‐T WAI‐T

correlation between the ANS‐T and the WAI‐T was run in order to WAI‐T .47 *
provide preliminary evidence for the construct validity of the measure. Feelings about client .26 ** .64 ***
The correlation between the two measures was moderate, at r = .468, Perceptions of work .25 * .70 ***
reflecting 21.9% shared variance and offering evidence of both con- Note. ANS‐T = Alliance Negotiation Scale–Therapist version; WAI‐
vergent and discriminant validity. Bivariate Pearson correlations dem- T = Working Alliance Inventory–Therapist version.
onstrated significant relationships between ANS‐T scores and how *Correlation significant at p < .05.
much therapists reported liking their clients overall, r = .26, p = .006, **Correlation significant at p < .01.
as well as how positively they felt about their work together, r = .25, ***Correlation significant at p < .001.
750 DORAN J. M. ET AL.

differences on ANS‐T scores. No significant differences emerged from the factor loadings of each item, which all fall above the recom-
across therapist or client demographic variables, including age, gender, mended benchmark for inclusion. The full scale meets all of the spec-
race/ethnicity, or sexual orientation (all ps > .05). Having therapeutic ified criteria outlined above and was determined to be both
dyads that were “matched” in terms of race/ethnicity (e.g., a White theoretically and statistically acceptable.
therapist with a White client) was not significantly different than those Evidence of construct validity is provided by the moderate corre-
that were “mismatched” (e.g., a White therapist with an African lation between the ANS‐T and WAI‐T. Although substantial overlap
American client, or vice versa), F(1, 112) = .12, p = .73, t = −1.60, between the client versions of the ANS and WAI initially raised some
p = .11. No diagnostic differences emerged as statistically significant, concerns about the distinction between the constructs and the utility
including the presence versus absence of a personality disorder, F(1, of a new alliance measure (Doran et al., 2016; Doran et al., 2017;
112) = 2.74, p = .10, t = .48, p = .64. Waizmann et al., 2015), convergence between the two measures
was lower in the present study. The ANS‐T and WAI‐T shared only
21.9% of variance in scores. The magnitude of the correlation
4 | DISCUSSION between the two measures (r = .468) offers important support for con-
vergent validity, demonstrating that the constructs are significantly
The present study was designed in order to create a therapist self‐ related to each other as would be theoretically expected. However,
report version of the ANS, previously only available in a client self‐ the correlation is also low enough to offer important evidence of dis-
report version. This paper presents the results of the initial scale criminant validity, showing that the two constructs are somewhat
development process. This research built on previous work (Doran operationally distinct.
et al., 2012) and involved a cross‐cultural collaboration with members In contrast with previous research on the ANS (Doran et al.,
of the research team behind the development of the Spanish‐language 2016), a lack of significant differences emerged in the current study
ANS‐A (Waizmann et al., 2015). The ANS and ANS‐A have demon- across demographic variables (both therapist and client), diagnostic
strated promise since their inception and have been met with interest variables, and treatment characteristics. Prior relationships of interest
from the psychotherapy research community. The absence of thera- have included the finding that negotiation is lower in clients with
pist versions of these measures has been limiting, and the present racial/ethnic minority identities, including in “mismatched” therapeutic
study aimed to address this gap, so that negotiation could begin to dyads, as well as in clients with self‐reported personality disorders.
be measured dyadically from the perspective of both client and These findings were not replicated in the present study, which may
therapist. be explained by the homogeneous nature of the sample (resulting in
The ANS‐T is similar to the original client version of the ANS, relatively small subgroups across categories), or the fact that data
though not an identical translation. Although the client version of were provided from the perspective of therapist rather than the client.
the ANS contains 12 items and two factors, the ANS‐T is unidimen- It is noteworthy that alliance negotiation, as measured by the ANS‐T,
sional and contains only nine items. Although it may have seemed demonstrated relationships with both therapist liking of their client
preferable to have a version of the ANS‐T that more closely mirrored and feelings about their work overall. Larger correlations emerged on
the client ANS (12 items and/or two factors), it was deemed more these same variables when using the WAI‐T, which indicate that the
important to create the most psychometrically sound scale possible. ANS‐T may offer more independence from these related constructs.
The resulting ANS‐T appears to offer a nice complement to the ANS. The ANS and ANS‐T are important and unique contributions to
Future research will be needed to determine how well these measures the psychotherapy research literature. They are very short, focused
complement and interact with each other, as well as to identify the measures designed to assess both client and therapist feelings, behav-
unique contributions of each. Another difference in the two scale iours, and perceptions of the therapeutic dyad. Their brief nature lends
development processes is that the ANS and the ANS‐A (Spanish ver- them to repeated measures assessment and reduces the burden of
sion of the client measure) were created sequentially, with the ANS measurement that can occur in complex psychotherapy studies. Ques-
developed in English and then tested as a translation in Spanish. In tions directly focus on reactions and responses to negative therapy
contrast, the ANS‐T and ANS‐TA (Gomez Penedo et al., 2017) were process, providing a measurement of the quality of the therapeutic
developed simultaneously. Of note, one of the three items that was relationship related to the ability to identify and address ruptures or
dropped from the ANS‐T was previously identified as statistically other negative therapy process when it occurs. The ANS includes
problematic in the initial ANS‐A validation study (Item 8: “I pretend more negatively valenced items than other alliance measures and
to agree with my therapist's goals for our therapy so the session runs focuses specifically on what happens when negative process occurs.
smoothly”). Developing the ANS‐T and ANS‐TA in tandem is advanta- This goes beyond traditional self‐report measures that may assess if
geous in that it does not privilege one culture over the other and has negative process exists but fail to capture what is done with it. There
resulted in two identical measures that are both psychometrically is broad consensus in the literature that negative therapy events such
strong in their respective populations. as ruptures are important aspects of the therapy process, and the suc-
In the present study, the ANS‐T emerged as a psychometrically cessful resolution of such events can be a catalyst for therapeutic
sound measure, with evidence for both the reliability and validity of change (Norcross & Wampold, 2011; Safran et al., 2011). Existing
the final version of the scale. An internal consistency analysis offered research on the presence of ruptures and their repair have tradition-
support for the reliability of the measure, despite the relatively small ally relied on observer‐based coding methods rather than client and
number of items comprising the scale. Content validity can be gleaned therapist self‐report (e.g., Eubanks‐Carter, Muran & Safran, 2015).
DORAN J. M. ET AL. 751

Although interesting and informative, such methods are very costly data from the therapy dyads under study. However, the major aim
and time consuming in nature. The ANS is the first measure that we of the study was to develop a therapist version of the scale for use
are aware of to assess the conditions conducive to the resolution of in future studies, and the feasibility of collecting dyadic data in large
alliance ruptures using a brief self‐report format. samples would have limited these efforts. Therefore, it was deter-
The ANS may also be an important tool in helping to further ongo- mined that future work should focus on analysing data from both cli-
ing work related to the concepts of sudden gains and sudden losses— ent and therapist measures. The current study also focused primarily
the appearance of abrupt therapeutic changes during therapy. Sudden on item and scale construction and preliminary reliability/validity anal-
gains refer to rapid improvement or indicators of significant change yses. The accumulation of validity evidence is a long‐term and ongoing
between sessions and have been found to be related to both short‐ process. Additional studies focused on assessing criterion and con-
term and long‐term treatment outcomes (Aderka, Nickerson, Bøe, & struct validity in more comprehensive ways will be necessary to fully
Hofmann, 2012; Stiles et al., 2003). Sudden losses refer to rapid evaluate the utility of the scale. The lack of clinical outcome data in
deteriorations during treatment and are associated with worse overall the current study is a limitation that will need to be addressed in
outcome (Lutz, Ehrlich, Rubel, Hallwachs, & Röttger, 2013). This phe- future studies, and the relationship between client‐ and therapist‐
nomenon has demonstrated some relationship to overall alliance levels ANS should be carefully examined. Furthermore, although the cross‐
and changes in the alliance (Lutz et al., 2013; Wucherpfennig, Rubel, cultural collaboration described above is primarily a strength of the
Hofmann, & Lutz, 2017). The ANS and ANS‐T have the potential to current study, the novelty of and lack of precedence for this approach
offer important insight about alliance negotiation quality and changes is nevertheless limiting. Finally, although factor analytic procedures
related to these constructs. For example, low alliance negotiation, or and recommendations were closely adhered to during data analysis,
sudden dips in alliance negotiation, may show a relationship to sudden any factor analytic procedure necessarily involves some subjective
losses in treatment. decision‐making, and which items to include or discard during scale
Despite decades of research focused on the relationship between development may be viewed as somewhat arbitrary in nature. To
session process, the therapeutic relationship, and treatment outcome address this, careful attention was given to adhering to the principles
in psychotherapy research (Lambert, 2013; Norcross, 2011), there of psychometric theory, and the authors tested and evaluated multiple
remains ample unexplained variance and critical gaps in our under- factor analytic solutions and configurations of the scale prior to
standing about what processes produce therapeutic change. Although selecting a final version. All scale items were subject to both rigorous
the client version of the ANS and ANS‐A have demonstrated some statistical and theoretical scrutiny during this multiphasic process.
promise in explaining incremental variance beyond what has been pre- Despite these limitations, the creation of a therapist version of the
viously accounted for by more traditional alliance measures (Doran ANS‐T is overdue and will be useful to clinicians and researchers inter-
et al., 2017; Roussos et al., 2017), more research is needed utilizing ested in the construct of alliance negotiation. Having a psychometri-
both client and therapist versions of the measure. Dyadic study of cally strong therapist measure to complement ongoing negotiation
negotiation is necessary in order to more fully understand the relation- research will add to the literature and better inform future investiga-
ship and the impact negotiation has on treatment outcome. Recent tions of negotiation and the ANS. Much more work will be needed
studies have demonstrated that convergence over time between cli- in order to further examine and validate the ANS‐T. The accumulation
ent‐ and therapist‐rated working alliances resulted in greater improve- of additional reliability and validity data will be needed to confirm the
ment of symptoms at termination (Coyne, Constantino, Laws, Westra psychometric integrity of the measure, and studies will also be needed
& Anthony, 2017; Laws et al., 2017). Whether increased convergence to examine its utility as a clinical and research tool. Given ongoing
in perceived negotiation also impacts treatment outcome will be cross‐cultural interest in the ANS, cross‐cultural research on the
important to examine in future studies. strengths and limitations of the measure in other languages and cul-
The present study has several limitations. First, in order to obtain tural contexts will be important. It is recommended that future studies
a sufficient sample for factor analytic techniques, online data collec- parallel existing work on the client versions of the ANS, investigating
tion methods were employed. Although online data collection has sev- the relationships between negotiation and both psychotherapy pro-
eral advantages, it also makes it impossible to confirm the accuracy of cess and outcome. Finally, the ANS and ANS‐T should be used
participant responses and may limit the representativeness and gener- together in subsequent studies going forward so that the relationship
alizability of the sample. Furthermore, it is possible that the results between them can also be analysed and understood.
may have been different if data were collected in person and/or
immediately following therapy sessions rather than on the therapists' CONFLIC T OF INT E RE ST
own time via an online survey. The study design was also cross‐ The authors have no conflicts of interest to report.
sectional in nature, and observations were limited to a single point in
time. It is possible that responses would have been different or more ORCID
variable if data had been collected using a repeated measures design
Jennifer M. Doran http://orcid.org/0000-0003-0921-4999
to allow for fluctuation in therapist perceptions. Although the sample
size was sufficient for the analyses that were conducted, it was never-
theless relatively small, and it would have been preferable to have had RE FE RE NC ES
a larger and more diverse sample. Another limitation is the use of only Aderka, I. M., Nickerson, A., Bøe, H. J., & Hofmann, S. G. (2012). Sudden
therapist self‐report data without collecting complementary client gains during psychological treatments of anxiety and depression: A
752 DORAN J. M. ET AL.

meta‐analysis. Journal of Consulting and Clinical Psychology, 80(1), 93– Horvath, A. O., & Greenberg, L. S. (1989). Development and validation of
101. https://doi.org/10.1037/a0026455 the Working Alliance Inventory. Journal of Counseling Psychology,
Aguirre McLaughlin, A., Keller, S. M., Feeny, N. C., Youngstrom, E. A., & 36(2), 223–233.
Zoellner, L. A. (2014). Patterns of therapeutic alliance: Rupture‐repair Horvath, A. O., Del Re, A. C., Fluckiger, C., & Symonds, D. (2011). Alliance
episodes in prolonged exposure for posttraumatic stress disorder. Jour- in individual psychotherapy. Psychotherapy: Theory, Research, Practice,
nal of Consulting and Clinical Psychology, 82, 112–121. https://doi.org/ Training, 48(1), 9–16.
10.1037/a0034696 Horvath, A. O., & Symonds, B. D. (1991). Relation between working alliance
Bachelor, A. (2013). Clients' and therapists' views of the therapeutic alli- and outcome in psychotherapy: A meta‐analysis. Journal of Counseling
ance: Similarities, differences, and relationship to therapy outcome. Psychology, 38, 139–149. https://doi.org/10.1037/0022‐0167.38.2.139
Clinical Psychology & Psychotherapy, 20, 118–135. https://doi.org/ Kendall, M. G., & Stuart, A. (1958). The advanced theory of statistics. (Vol. 1).
10.1002/cpp.792 London: Griffin.
Bordin, E. S. (1979). The generalizability of the psychoanalytic concept of Kline, P. (1994). An easy guide to factor analysis. London: Routledge.
the working alliance. Psychotherapy: Theory, Research, and Practice, 16,
252–260. https://doi.org/10.1037/h0085885 Lambert, M. J. (Ed.) (2013). Bergin & Garfield's handbook of psychotherapy and
behavior change (sixth ed.). Hoboken, New Jersey: John Wiley & Sons.
Bryant, F. B., & Yarnold, P. R. (1995). Principal components analysis and
exploratory and confirmatory factor analysis. In L. G. Grimm, & P. R. Laws, H. B., Constantino, M. J., Sayer, A. G., Klein, D. N., Kocsis, H.,
Yarnold (Eds.), Reading and understanding multivariate statistics. Manber, R., … Arnow, B. A. (2017). Convergence in patient–therapist
Washington DC: American Psychological Association Books. therapeutic alliance ratings and its relation to outcome in chronic
depression treatment. Psychotherapy Research, 27(4), 410–424.
Clark, L. A., & Watson, D. (1995). Constructing validity: Basic issues in https://doi.org/10.1080/10503307.2015.1114687
objective scale development. Psychological Assessment, 7(3), 309–319.
Lutz, W., Ehrlich, T., Rubel, J., Hallwachs, N., & Röttger, M. (2013). The ups
Coyne, A. E., Constantino, M. J., Laws, H. B., Westra, H. A., & Antony, M. and downs of psychotherapy: Sudden gains and sudden losses identi-
M. (2017). Patient–therapist convergence in alliance ratings as a pre- fied with session reports. Psychotherapy Research, 23, 37–41. https://
dictor of outcome in psychotherapy for generalized anxiety disorder. doi.org/10.1080/10503307.2012.693837
Psychotherapy Research. 1–16. Online First Publication. https://doi.
org/10.1080/10503307.2017.1303209 Mander, J., Schlarb, A., Teufel, M., Keller, F., Hautzinger, M., Zipfel, S., …
Sammet, I. (2015). The Individual Therapy Process Questionnaire:
Cushman, P., & Gilford, P. (2000). Will managed care change our way of Development and validation of a revised measure to evaluate general
being? American Psychologist, 55(9), 985–996. https://doi.org/ change mechanisms in psychotherapy. Clinical Psychology & Psychother-
10.1037/0003‐066X.55.9.985 apy, 22(4), 328–345.
Doran, J. M. (2016). The working alliance: Where have we been, where are Mitchell, S., & Aron, L. (1999). Relational psychoanalysis: The emergence of a
we going? Psychotherapy Research, 26(2), 146–163. https://doi.org/ tradition. Hillsdale, NJ: The Analytic Press.
10.1080/10503307.2014.954153
Norcross, J. C. (Ed.) (2011). Psychotherapy relationships that work: Evidence‐
Doran, J. M., Safran, J. D., & Muran, J. C. (2016). The Alliance Negotiation based responsiveness (second ed.). New York: Oxford University Press.
Scale: A psychometric investigation. Psychological Assessment, 28(8),
885–897. https://doi.org/10.1037/pas0000222 Norcross, J. C., & Wampold, B. E. (2011). Evidence‐based therapy relation-
ships: Research conclusions and clinical practices. In J. C. Norcross
Doran, J. M., Safran, J. D., & Muran, J. C. (2017). An investigation of the (Ed.), Psychotherapy relationships that work: Evidence‐based responsive-
relationship between the Alliance Negotiation Scale and psychother- ness (2nd ed.) (pp. 423–430). New York: Oxford University Press.
apy process and outcome. Journal of Clinical Psychology, 73(4), 449–
465. https://doi.org/10.1002/jclp.22340 Piper, W. E., Azim, H. F. A., Joyce, A. S., & McCallum, M. (1991). Transfer-
ence interpretations, therapeutic alliance and outcome in short‐term
Doran, J. M., Safran, J. D., Waizmann, V., Bolger, K., & Muran, J. C. (2012). individual therapy. Archives of General Psychiatry, 48, 946–953.
The Alliance Negotiation Scale: Psychometric construction and prelim- https://doi.org/10.1001/archpsyc.1991.01810340078010
inary reliability and validity analysis. Psychotherapy Research, 22(6),
710–719. Roussos, A., Gomez Penedo, J. M., Doran, J. M., Olivera, J., & Horowitz, L.
(2017). Early alliance negotiation as a predictor of therapeutic change
Eubanks, C.F., Muran, J.C. & Safran, J.D. (2015). Rupture resolution rating in psychotherapy. Journal of Clinical Psychology. Under review
system (3RS): Manual. Unpublished manuscript, Mount Sinai‐Beth
Israel Medical Center, New York. Safran, J. D. (1993). Breaches in the therapeutic alliance: An arena for
negotiating authentic relatedness. Psychotherapy, 30(1), 11–24.
Gelso, C. J., Kelley, F. A., Fuertes, J. N., Marmarosh, C., Holmes, S. E., & https://doi.org/10.1037/0033‐3204.30.1.11
Costas, C. R. (2005). Measuring the real relationship in psychotherapy:
Initial validation of the therapist form. Journal of Counseling Psychology, Safran, J. D., Crocker, P., McMain, S., & Murray, P. (1990). Therapeutic
52, 640–649. https://doi.org/10.1037/0022‐0167.52.4.640 alliance rupture as a therapy event for empirical investigation. Psycho-
therapy, 27, 154–165. https://doi.org/10.1037/0033‐3204.27.2.154
Gomez Penedo, J.M. (2017). Variables del paciente y del proceso
psicoterapéutico vinculadas a predictores tempranos de cambio. Safran, J. D., & Kraus, J. (2014). Alliance ruptures, impasses, and enact-
Doctoral dissertation, Universidad de Buenos Aires, Buenos Aires, ments: A relational perspective. Psychotherapy, 51, 381–387.
Argentina. Safran, J. D., & Muran, J. C. (2000). Negotiating the therapeutic alliance: A
Gomez Penedo, J.M., Doran, J.M. & Roussos, A. (2017). Perspectiva del relational treatment guide. New York: Guilford Press.
terapeuta acerca de la negociación de la alianza: Un instrumento para Safran, J. D., & Muran, J. C. (2006). Has the concept of the alliance outlived
su evaluación. Under review. its usefulness? Psychotherapy, 43, 286–291.
Gregory, R. J. (2004). Psychological tests: History, principles, and applications Safran, J. D., Muran, J. C., & Eubanks‐Carter, C. (2011). Repairing alliance
(4th ed.). Boston, MA: Pearson Education. ruptures. Psychotherapy, 48, 80–87. https://doi.org/10.1037/a0022140
Hersoug, A. G., Hoglend, P., Monsen, J. T., & Havik, O. E. (2001). Quality of Stiles, W. B., Glick, M. J., Osatuke, K., Hardy, G. E., Shapiro, D. A., Agnew‐
working alliance in psychotherapy. The Journal of Psychotherapy Prac- Davies, R., … Barkham, M. (2004). Patterns of alliance development and
tice and Research, 10, 205–216. the rupture‐repair hypothesis: Are productive relationships U‐shaped
Horvath, A. O., & Greenberg, L. S. (1986). The development of the Working or V‐shaped? Journal of Counseling Psychology, 51, 81–92. https://doi.
Alliance Inventory. In L. S. Greenberg, & W. M. Pinsof (Eds.), The psy- org/10.1037/0022‐0167.51.1.81
chotherapeutic process: A research handbook (pp. 529–556). New York: Stiles, W. B., Leach, C., Barkham, M., Lucock, M., Iveson, S., Shapiro, D. A.,
The Guildford Press. … Hardy, G. E. (2003). Early sudden gains in psychotherapy under
DORAN J. M. ET AL. 753

routine clinic conditions: Practice‐based evidence. Journal of Consulting Waizmann, V., Doran, J. M., Bolger, K., Gomez Penedo, J. M., Safran, J. D.,
and Clinical Psychology, 71, 14–21. https://doi.org/10.1037/0022‐ & Roussos, A. J. (2015). Escala de Negociación de Alianza Terapéutica
006X.71.1.14 (ANS‐A): Evidencias de su validez y confiabilidad. Revisita Argentina
de Clínica Psicológica, XXIV(3), 243–254.
Strauss, J. L., Hayes, A. M., Johnson, S. L., Newman, C. F., Brown, G. K.,
Barber, J. P., … Beck, A. T. (2006). Early alliance, alliance ruptures, and Wucherpfennig, F., Rubel, J. A., Hofmann, S. G., & Lutz, W. (2017). Pro-
symptom change in a nonrandomized trial of cognitive therapy for cesses of change after a sudden gain and relation to treatment
avoidant and obsessive‐compulsive personality disorders. Journal of outcome: Evidence for an upward spiral. Journal of Consulting and Clin-
Consulting and Clinical Psychology, 74, 337–345. https://doi.org/ ical Psychology, 85, 1199–1210. https://doi.org/10.1037/ccp0000263
10.1037/0022‐006X.74.2.337

Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics. Boston,

How to cite this article: Doran JM, Gómez‐Penedo JM,
MA: Allyn and Bacon.
Safran JD, Roussos A. A therapist version of the Alliance Nego-
Tichenor, V. U., & Hill, C. E. (1989). A comparison of six measures of the
tiation Scale. Clin Psychol Psychother. 2018;25:745–753.
working alliance. Psychotherapy, 26, 195–199. https://doi.org/
10.1037/h0085419 https://doi.org/10.1002/cpp.2197

Tracey, T. J., & Kokotovic, A. M. (1989). Factor structure of the working

alliance inventory. Psychological Assessment, 1(3), 207–210. https://
doi.org/10.1037/1040‐3590.1.3.207

APPENDIX A.
The Alliance Negotiation Scale–Client version
Doran et al. (2012)

Please answer the following questions based on how you feel with your therapist overall.
Never Sometimes Always

1. I am comfortable expressing frustration with my therapist when it arises. 1 2 3 4 5 6 7

2. I feel that I can disagree with my therapist without harming our relationship. 1 2 3 4 5 6 7
3. My therapist encourages me to express any concerns I have with our progress. 1 2 3 4 5 6 7
4. My therapist and I are not good at finding a solution if we disagree. 1 2 3 4 5 6 7
5. My therapist is inflexible and does not take my wants or needs into consideration. 1 2 3 4 5 6 7
6. I am comfortable expressing disappointment in my therapist when it arises. 1 2 3 4 5 6 7
7. My therapist encourages me to express any anger I feel towards him/her. 1 2 3 4 5 6 7
8. I feel like I do not have a say regarding what we do in therapy. 1 2 3 4 5 6 7
9. I feel that my therapist tells me what to do, without much regard for my wants or needs. 1 2 3 4 5 6 7
10. I pretend to agree with my therapist's goals for our therapy so the session runs smoothly. 1 2 3 4 5 6 7
11. My therapist is rigid in his/her ideas regarding what we do in therapy. 1 2 3 4 5 6 7
12. My therapist is able to admit when he/she is wrong about something we disagree on. 1 2 3 4 5 6 7

Note. The measure was first published in Doran et al., 2012

A P P E N D I X B.

The Alliance Negotiation Scale–Therapist version

Please answer the following questions based on how you feel with your patient overall.
Never Sometimes Always

1. I feel that I can disagree with my patient without harming our relationship. 1 2 3 4 5 6 7
2. I am comfortable hearing my patient's negative feelings about me or our work. 1 2 3 4 5 6 7
3. I believe my patient feels comfortable expressing disappointment in me when it arises. 1 2 3 4 5 6 7
4. I encourage my patient to express any anger he/she feels towards me. 1 2 3 4 5 6 7
5. I am able to admit to my patient when I am wrong about something we disagree on. 1 2 3 4 5 6 7
6. I believe my patient feels comfortable expressing frustration in me when it arises. 1 2 3 4 5 6 7
7. My patient and I are able to constructively work through and resolve tension or ruptures in our relationship. 1 2 3 4 5 6 7
8. I believe my patient feels like he/she has a say regarding what we do in therapy. 1 2 3 4 5 6 7
9. I regularly “check in” with my patient to see if he/she feels that the way we are working together is correct 1 2 3 4 5 6 7
Copyright of Clinical Psychology & Psychotherapy is the property of John Wiley & Sons,
Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv
without the copyright holder's express written permission. However, users may print,
download, or email articles for individual use.