See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/262110711

The impact of early empathy on alliance building, emotional processing, and

outcome during experiential treatment of depression

Article in Psychotherapy Research · May 2014

DOI: 10.1080/10503307.2014.901572 · Source: PubMed

CITATIONS READS

61 754

2 authors:

Ashley J Malin Alberta Pos

University of Florida York University
33 PUBLICATIONS 618 CITATIONS 23 PUBLICATIONS 1,000 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Alberta Pos on 29 May 2014.

The user has requested enhancement of the downloaded file.

This article was downloaded by: [Alberta Pos]
On: 09 May 2014, At: 09:12
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Psychotherapy Research
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/tpsr20

The impact of early empathy on alliance building,

emotional processing, and outcome during experiential
treatment of depression
a a
Ashley J. Malin & Alberta E. Pos
a
Department of Psychology, York University, Toronto, ON, Canada
Published online: 07 May 2014.

To cite this article: Ashley J. Malin & Alberta E. Pos (2014): The impact of early empathy on alliance building,
emotional processing, and outcome during experiential treatment of depression, Psychotherapy Research, DOI:
10.1080/10503307.2014.901572

To link to this article: http://dx.doi.org/10.1080/10503307.2014.901572

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Psychotherapy Research, 2014
http://dx.doi.org/10.1080/10503307.2014.901572

EMPIRICAL PAPER

The impact of early empathy on alliance building, emotional processing,

and outcome during experiential treatment of depression

ASHLEY J. MALIN & ALBERTA E. POS

Department of Psychology, York University, Toronto, ON, Canada

(Received 14 February 2013; revised 28 February 2014; accepted 3 March 2014)

Abstract
Objective: This study examined the relationships among the therapist process of expressed empathy during first sessions,
clients’ post-session one alliance reports, clients’ later working phase emotional processing, and clients’ final reductions in
Downloaded by [Alberta Pos] at 09:12 09 May 2014

depressive symptoms for 30 clients receiving short-term experiential therapy for depression. Method: The therapist process
of expressed empathy was assessed using a new observer-rated measure: the measure of expressed empathy, which was
demonstrated to be valid and reliable. Results: Results indicate that therapist expressed empathy in session one significantly
affected the outcome, albeit indirectly. This indirect effect occurred through two direct effects on other important therapy
processes that did directly predict client outcomes: (i) Therapist expressed empathy in first sessions directly and positively
predicted client reports of first-session alliances; and (ii) therapist expressed empathy directly predicted observer-rated
deepened client emotional processing in the working phase of therapy. Conclusions: Empirical support was provided for
the theorized relationships in experiential theory amongst the variables examined.

Keywords: empathy; working alliance; emotional processing; experiential therapy; depression

This study investigated the importance of the pro-
cess of therapist expressed empathy during first
sessions of experiential therapy to both small “o”
(session processes) and large “O” (therapy) out-
comes (Greenberg, 1986) during experiential treat-
ment of depression. The relationships among
therapist expressed empathy, clients’ session one
reports of the alliance, clients’ working phase emo-
tional processing (WPEP), and clients’ post-therapy
reports of reductions in depressive symptoms were
explored.
Depression is currently the leading cause of
disease burden in high-income countries, and is
forecast to become the leading cause of disease
burden worldwide by 2030 (Ferrari et al., 2013;
World Health Organization, 2004). While interper-
sonal, cognitive-behavioural, and experiential psy-
chotherapies are all empirically validated and equally
effective short-term treatments for moderately
depressed adults (Elkin et al., 1989; Goldman,
Greenberg, & Angus, 2006; Greenberg & Watson,
1998; Shea & Elkin, 1996; Watson, Gordon, Ster-
mac, Kalogerakos, & Steckley, 2003), their effects
can be considered modest (Westen & Morrison,
2001). As such, all treatments should be improved.
To accomplish this, researchers must establish how
these treatments work.
It is often currently argued that after client factors
(Lambert, 1992), common, rather than specific,
factors are responsible for the lion’s share of client
change in all psychotherapies. Empathy and the
working alliance are two suggested common factors
(see Norcross, 2011; Greenberg & Watson, 2006).
Emotional processing (EP) has been suggested as a
third common factor across all treatments for
depression (Greenberg & Pascual-Leone, 2006 for
a review; Pos, Greenberg, & Warwar, 2009; Teas-
dale, 1999). Common factors, however, are specif-
ically implemented within different models of
treatment (Goldfried, 2004). Therefore, the pre-
sumption of common factors does not excuse
researchers from the task of empirically validating

Correspondence concerning this article should be addressed to Ashley J. Malin, Department of Psychology, York University, 4700 Keele
Street Toronto, ON M3J 1P3, Canada. Email: ashleyjs@yorku.ca

© 2014 Society for Psychotherapy Research

 2 A. J. Malin and A. E. Pos
mechanisms of change within particular effective
treatments. Wampold (2001) has suggested that
this requires us to demonstrate that a theoretically
informed implementation of a particular common
factor (such as alliance development) within a spe-
cific model of treatment (such as experiential
therapy) occurs and can predict client improvement.
An additional concern, particularly related to
research on the alliance, is the frequent use of
transtheoretical measures, such as the Working
Alliance Inventory (WAI; Horvath & Greenberg,
1989), to operationalize the alliance. This often
causes us to ignore exploring variables contributing
to the development of the alliance within particular
approaches. Horvath and Bedi (2002) have sug-
gested attention be paid to these factors that influ-
ence alliance development, especially what the
therapist does in session (therapist process) to
predict good alliance formation. This study tested
experiential theory concerning change, in particular
the importance placed on therapists expressing
Downloaded by [Alberta Pos] at 09:12 09 May 2014
empathy to alliance building, facilitating EP, and
final outcomes, during experiential therapy for
depression.
Empathy
Empathy is a basic relational process (Elliott &
Greenberg, 2007; Goleman, 1995; Rogers, 1959)
and an important component of emotional intelli-
gence. It is also a core relationship condition experi-
ential therapists employ to establish the alliance and
to support EP (Pos, Greenberg, & Elliott, 2008).
Defined by Rogers (1959) as the ability to accurately
perceive the internal frame of reference of another,
including their emotional components and mean-
ings, empathy is a complex higher-order construct.
This means that therapists must utilize a full range of
cognitive and emotional functions while responding
empathically to clients. Therapist behaviours shown
to significantly relate to client perceptions of
empathy include: Therapists’ non-verbal behaviours
[maintaining eye contact, having a concerned
expression, and maintaining a forward lean or head
nodding to convey an understanding (D’Augelli,
1974; Tepper, 1973; Watson, 2001)]; therapists’
speech characteristics [having similar rates of speech
and vocal tones as their clients, responding just
ahead of their clients, and not interrupting (Barring-
ton, 1961; Elliott, Bohart, Watson, & Greenberg,
2011; Greenberg & Elliott, 1997)]; therapists’
response modes [conveying a sense of interest,
having an equal level of emotional involvement as
their clients, not conveying detachment or boredom
(Caracena & Vicory, 1969; Tepper, 1973), use of
emotion words (Barrington, 1961), provision of
exploratory responses rather than general advice
(Barkham & Shapiro, 1986), and clearly commu-
nicated messages (Bohart & Greenberg, 1997;
Caracena & Victory, 1969)]. Therapists’ character-
istics such as being non-judgemental, attentive, and
open to discussing any topic also have been shown to
be important (Myers, 2000).
Empathy and Alliance Formation
Experiential theory assumes that therapists’ provi-
sion of empathy is the foundation upon which the
alliance is constructed. Experiential therapists are
directed from the first moments of therapy to engage
in and provide an empathically attuned relationship
for the purpose of promoting safety and facilitating
alliance development (Greenberg & Watson, 2006;
Pos et al., 2008; Rogers, 1975). In fact, significant
positive associations between empathy and the work-
ing alliance have been found across a number of
therapeutic approaches. Salvio, Beutler, Wood, and
Engle (1992) found moderate-to-large correlations
between clients’ ratings of empathy and the alliance,
both measured after session 20 for clients receiving
various types of therapy for depression. Horvath
(1981) and Mosely (1983) both found moderate-
to-large correlations between clients’ ratings of
empathy and the alliance after the third session of
various types of therapy. Finally, Wing (2010) found
a significant moderate correlation between observer
ratings of therapist empathy and alliance formation
for clients receiving psychodynamic psychotherapy.
The present study, however, tests experiential theory
concerning the particular importance of first-session
therapist empathy to initial alliance building or later
therapy processes and outcome.
The importance of first-session alliances during
experiential therapy for depression has been high-
lighted by Pos et al. (2009) who found that even after
controlling for growth in the alliance and deepening
of EP across therapy, clients’ reports of the alliance
after first sessions directly predicted all outcome
measures. Therefore, alliance-related factors operat-
ing within the very first hour of experiential therapy
appear to be important for clients’ final outcomes. A
long-held assumption that three to five sessions are
required to develop an alliance has resulted in the
majority of alliance research not adequately examin-
ing alliance development in initial sessions nor the
potential ramification of first-session alliances to later
processes and final outcomes (see Horvath & Bedi,
2002 for a review). Support for examining first-session
alliances is offered by dynamic systems theory that
argues that initial conditions within any process set the
course for its later developmental trajectory (van
Geert, 1996). Furthermore, experiential theory

(Greenberg & Watson, 2006) and recent experiential
process research (Pos et al., 2009; Wong & Pos, 2014)
suggest that alliance building in the initial hour can
considerably influence therapy process and outcome.
Therefore, for both empirical and theoretical reasons
the importance of examining first sessions and
empathic process in first sessions is underlined.
Empathy and EP
Experiential theory also assumes that EP is a core
change process in experiential therapy (Pascual-
Leone & Greenberg, 2007; Pos, Greenberg, Gold-
man, & Korman, 2003; Pos et al., 2009). Within
experiential psychotherapy, adaptive EP is seen as
involving an integration of cognitive and affective
components (Greenberg, 2002) and consisting of a
series of stages (Pos et al., 2003; Pos & Greenberg,
2007). Clients learn to approach, be aware of and
tolerate contact with emotions, cognitively orient to
and make meaning of emotional information, and
Downloaded by [Alberta Pos] at 09:12 09 May 2014
finally transform emotion by accessing adaptive
emotional resources (Pos & Greenberg, 2007).
Through this process, new emotional reactions and
meanings can emerge that can be integrated into
cognitive–affective meaning structures and subse-
quently change them (Pos, Greenberg, & Goldman,
2005). Experiential therapists are taught that provid-
ing an empathically attuned relationship from the
start of therapy will facilitate clients’ EP by helping
them feel safe enough to move their attention away
from interpersonal concerns towards engaging fully
in the therapy task of approaching their feelings (Pos,
2006; Watson, Goldman, & Vanaerschot, 1998).
Developmental researchers also agree that therapist
empathy can facilitate EP and believe that it does so
by helping to heal empathic failures that were
experienced during childhood (Paivio & Laurent,
2001). Although the relationship between therapist
empathy and later EP is assumed to exist, it has not
yet been empirically tested. Some evidence of this
relationship was found by Steckley (2006) who
demonstrated that clients who perceived their thera-
pists as being empathic experienced structural
changes in their internal models of self and other,
and treated themselves and others less negatively
(were less destructive, controlling, oppressing, crit-
ical, rejecting, careless, and neglecting) by the end of
therapy. All of these positive self-changes were also
associated with more favourable treatment outcomes
(Steckley, 2006).
Empathy and Outcome
The alliance and EP have already been demonstrated
to be robust predictors of psychotherapy outcome in
Psychotherapy Research 3
experiential therapy for depression (Pos, Greenberg,
Goldman, & Korman, 2005; Pos et al., 2009).
Empathy also has been demonstrated to be a moder-
ately strong predictor of and to account for approxi-
mately 9% of the variance in psychotherapy outcome
across many models of intervention (Elliott et al.,
2011). Furthermore, empathy has been demon-
strated to account for as much or more variance in
outcome than specific psychotherapeutic interven-
tions (Bohart, Elliott, Greenberg, & Watson, 2002).
Other variables such as expressing understanding,
being attentive, and being openly receptive to clients’
perspectives have also been associated with success-
ful outcomes (Henry, Schact, & Strupp, 1986;
Watson, Enright, & Kalogerakos, 1998) and are
identified as important to perceived empathy.
Empathy Measures
Therapist empathy has been measured by client
reports, therapist reports, assessing therapist and
client perceptual congruence of therapist empathy,
or by observer ratings of expressed empathy (Elliott
et al., 2011; Watson & Prosser, 1999/2002).
Research has consistently shown that client-rated
measures are the best predictors of psychotherapy
outcome, followed closely by observer-rated mea-
sures (Elliot et al., 2011; Watson & Prosser, 1999/
2002), with therapist-rated measures being the least
predictive (Bohart et al., 2002; Elliott et al., 2011).
Degree of therapist and client perceptual congruence
has been found to be unrelated to psychotherapy
outcome (Elliott et al., 2011). The Barrett-Lennard
Relationship Inventory (BLRI; Barrett-Lennard,
1986) is the best known and most widely used
client-rated measure of therapist empathy that
assesses clients’ perceptions of empathy, as opera-
tionalized by Rogers (1957; Elliott et al., 2011).
Using this measure, clients rate the extent to which
they experience the therapist as genuine, prizing, or
empathic during the therapy session. Two early
observer-rated empathy measures designed by Truax
and Carkhuff (1967) and Carkhuff and Berenson
(1967) also widely used in the past have been
criticized for not adequately reflecting client-centred
(CC) conceptions of empathy as an attitude and for
focusing exclusively on empathic reflections (Lam-
bert, De Julio, & Stein, 1978; Watson & Prosser,
1999/2002). In the present study, we employed a
relatively new observer-rated measure: The measure
of expressed empathy (MEE; Watson & Prosser,
1999/2002). It differs from other measures in that it
is based on the behavioural correlates of empathy
identified in previous research (Barrington, 1961;
Bohart & Greenberg, 1997; Caracena & Victory,
1969; D’Augelli, 1974; Elliott et al., 2011;
 4 A. J. Malin and A. E. Pos
Greenberg & Elliott, 1997; Tepper, 1973; Watson,
2001) and evaluates both therapists’ verbal and non-
verbal behaviour, including speech characteristics
and response modes. Unlike client-rated measures,
such as the BLRI that assesses clients’ global
perceptions of empathy, the MEE measures specific
components of expressed empathy (Watson & Pros-
ser, 1999/2002). Due also to the fact that we were
interested in testing theoretical assumptions con-
cerning observable (not perceived) therapist beha-
viours that relate to later client reports and
experiences, the MEE was assumed to be an optimal
measure of therapists engaging in the process of
expressing empathy.
Current Study
The hypotheses of the current study are summarized
as follows: (i) Since the experiential assumption is
that therapist empathy in session one is a key
Downloaded by [Alberta Pos] at 09:12 09 May 2014
therapist process, the first hypothesis is that therapist
expressed empathy in session one will positively
predict post-session one client-rated alliance scores.
(ii) Another experiential assumption is that therapist
empathy is directly essential to facilitating EP.
Therefore, hypothesis two is that therapist expressed
empathy in first sessions will positively predict EP in
the working phase of therapy. Finally, our last
hypothesis concerns empathy’s relationship to out-
come. Given that therapist empathy in session one is
assumed to facilitate two important change processes
(i.e., the formation of the working alliance and later
EP) that predict outcome in this therapeutic modal-
ity, and that therapist empathy measured later on in
therapy has been found to be predictive of outcome,
we expected that therapist expressed empathy meas-
ured in session one would directly predict outcome
as well.
Research Design and Method
Participants
Due to the exploratory nature of the present study,
the sample consisted of 30 clients with the lowest
(LA) and highest post-session one alliance (HA)
ratings selected from a larger sample of 74 clients
who received short-term (16–20 sessions) treatment
in two York University trials of experiential psycho-
therapy for depression (Goldman et al., 2006;
Greenberg & Watson, 1998). Both trials compared
the effectiveness of two experiential treatments: CC
therapy and emotion-focused therapy (EFT). All
subjects met criteria for major depressive disorder
on the Structured Clinical Interview (SCI; Spitzer,
Williams, Gibon, & First, 1989) for the Diagnostic
and Statistical Manual of Mental Disorders
(DSM-III-R; APA, 1987). All 30 clients completed
treatment. For information on the original screening,
assessment, and treatment procedures, see Green-
berg and Watson (1998) and Goldman et al. (2006).
The inclusion criteria were: A score of 16 or
greater on the Beck Depression Inventory (BDI-LF;
Beck, Ward, Mendelson, Mock, & Erbaugh, 1961),
a score greater than 50 on the Global Assessment of
Functioning on the DSM-III-R, and being between
the ages of 18 and 65 at the time of assessment.
Exclusion criteria were: Current drug or alcohol
abuse; bipolar or psychotic disorder, current eating
disorder, antisocial or borderline personality dis-
order; recent suicide attempts; a past history of
incest; the loss of a significant other in the past
year; or involvement in an ongoing violent relation-
ship. These criteria assured that severely depressed,
functionally impaired subjects were excluded (Elkin
et al., 1989).
Clients (n = 18) in the HA group ranged in age
from 26 to 55 (M = 38.89, SD = 9.88) and clients in
the LA group (n =12) ranged in age from 31 to 63
(M = 48.92, SD = 10.22). Clients in the HA group
were 75% (n = 9) female and 25% (n = 3) male,
while those in the LA group were 72.2% (n = 13)
female and 27.8% (n = 5) male. No significant
differences were found between HA and LA groups
in terms of marital status, gender, or type of therapy
received (CC or EFT). Significant differences
between HA and LA groups in age and education
status were found. Clients in the HA group were
younger (M = 38.89, SD = 9.88) versus the LA
group (M = 48.92, SD = 10.22), p = .01, and tended
to have higher education statuses than those in the
LA group, p < .05. However, since age or education
did not predict any therapy process measures they
were not included as predictors in any analyses.
Therapists
There were 16 therapists in this study: 4 males and 12
females. Twelve were clinical psychology doctoral
students (2 male, 10 female), two (one male and one
female) were psychiatrists, and three (two male and
one female) were psychologists. Two therapists had
clients in both HA and LA groups, while eight
therapists had clients in the HA group only and six
therapists had clients in the LA group only. Therapists
received 40 hr of manual-based training, supervision
by licensed psychologists, and adherence monitoring.
Treatments
Two experiential therapies were used: CC therapy
(Rice, Greenberg, & Watson, 1994; Rogers, 1957,

1975) and EFT (Greenberg, Rice, & Elliott, 1993;
Greenberg & Watson, 2006). Clients were randomly
assigned to treatment. For both approaches, the
primary goal during the first three sessions is for
therapists to provide empathic attunement to pro-
mote the formation of a strong therapeutic alliance
and facilitate a deepening of clients’ EP (Pos et al.,
2008). Thus, therapy process does not differ
between CC and EFT within the first three sessions.
The core goal of both therapies is to deepen the
client’s emotional and experiential processing.
Adherence was measured and achieved in both con-
ditions (see the original outcome reports Goldman
et al., 2006; Greenberg & Watson, 1998).
CC therapy. This therapy followed a manual
based on Rogers (1957, 1975; Rice et al., 1994). In
CC therapy, therapists provide the necessary facil-
itative relationship conditions: Unconditional posit-
ive regard, empathy, and congruence. Following the
clients’ internal track, therapists communicate
Downloaded by [Alberta Pos] at 09:12 09 May 2014
empathy, facilitate exploration, encourage symbol-
ization of core meaning, and increase emotional
awareness.
Emotion-focused therapy. In this therapy,
therapists, after session three and within the context
of a CC relationship, use marker-guided, process-
directive interventions intended to promote optimal
EP. Specific emotional problem markers are used to
determine matched interventions. Four interventions
used in the study were: Two-chair dialogue for self-
evaluative conflict; empty-chair dialogue for unfin-
ished business with a significant other; focusing on
an unclear felt sense (Gendlin, 1997); and systematic
evocative unfolding for problematic or confusing (to
the client) affective reactions. Currently, there are
two manuals for the use of this therapy for depres-
sion (Greenberg et al., 1993; and Greenberg &
Watson, 2006).
Process Measures
Measure of expressed empathy. This is an
observer-rated measure of therapist-communicated
empathy that evaluates therapists’ verbal and non-
verbal behaviours, speech characteristics, and
response modes (Watson & Prosser, 1999/2002). It
consists of 10 dimensions that are rated on a 9-point
Likert scale (from 0 = “never” to 8 = “all the time”)
based on the percentage of time that the behaviour
was present during the rated segment. A global
empathy score is calculated as the average of the 10
subscale ratings. Internal consistency for the scale, as
assessed by scale developers, is high (α = .88).
Construct validity is provided by a large and
Psychotherapy Research 5
significant correlation (r = .66, p < .01) with client
ratings of empathy, as measured by the Barrett-
Lennard Relationship Inventory (BLRI; Barrett-
Lennard, 1962), a well-validated client-rated measure
of empathy, and the fact that it is based on empirically
validated behavioural predictors of empathy.
The 10 therapist expressive dimensions assessed
by the MEE are: (i) Vocal concern: The extent to
which the therapist’s voice has a soft resonance with
a grounded open quality; (ii) vocal expressiveness:
The extent to which the expressiveness in the
therapist’s voice varies appropriately in energy, col-
our, and pitch to respond to the nature of the client’s
subject matter; (iii) vocal matching: The degree to
which the therapist’s vocal response quality appro-
priately matches and/or holds the intensity of emo-
tion that the client is experiencing; (iv) warmth and
interpersonal safety: The degree to which the ther-
apist communicates an atmosphere of warm safety
through soft expressiveness, smiling, and eye con-
tact; (v) responsive attunement: The extent to which
the therapist tracks and responds to the client’s
moment-to-moment experience (including facial and/
or non-verbal behaviours that may differ from verbal
content); (vi) look of concern: The extent to which the
therapist appears caring, engaged, and involved (not
bored or disinterested); (vii) responsiveness/following:
The extent to which the therapist adjusts responses
to follow the client’s track rather than lead session
content; (viii) understanding experience: The degree
to which the therapist generally communicates sens-
itive understanding of the significance of the client’s
inner world of emotional meaning relating to events
being discussed; (ix) understanding cognitive mean-
ing: The degree to which the therapist conveys
accurate understanding of the client’s cognitive
meaning framework by following and understanding
both their client’s narratives as well as the client’s
idiosyncratic construals explicit or implicit in these
narratives; (x) therapist genuineness/acceptance: The
extent to which the therapist communicates that they
value and prize the client and appears sincere,
authentic, and genuine (Watson & Prosser,
1999/2002).
Working alliance inventory. This is a 36-item
measure of the therapeutic alliance rated on a
7-point Likert scale composed of three subscales
that each assesses the client–therapist bond, client–
therapist agreement on the tasks, and goals of
therapy (Horvath & Greenberg, 1986, 1989). The
internal consistency of the entire scale is reportedly
high (.87–.93) as is it for the individual subscales
(.92 for Bond, .90 for Task, and .89 for Goal;
Horvath & Greenberg, 1986, 1989). There is a
short-form (12-item) version with comparable
 6 A. J. Malin and A. E. Pos
psychometric properties (Tracey & Kokotovic,
1989). Client-rated short forms were used in this
study.
The experiencing scale. This scale measures the
degree to which clients orient to, symbolize, and use
internal felt experience to inform problem–solving
(Klein, Mathieu-Coughlan, & Kiesler, 1986). Rat-
ings are given on a 7-point ordinal rating scale and
are assigned to segments of psychotherapy based on
grammatical, expressive, paralinguistic, and content
distinctions indicative of different degrees of experi-
encing. Ratings from 1 to 4 describe the progressive
movement of orientation from external to internal
referents, while ratings from 5 to 7 describe the
progressive use of experienced inner perspectives in
affective problem-solving. Inter-rater reliability coef-
ficients have been reported to range from .76 to .91,
and rating or re-rating coefficients of .80 have been
reported (Pos et al., 2009). This scale was applied to
Downloaded by [Alberta Pos] at 09:12 09 May 2014
emotion episodes (EEs; narrative segments within
which clients described real or imagined emotional
experiences; see Pos et al., 2009) in this study.
Session Outcome Measures
General session evaluation questionnaire.
The General Session Evaluation Questionnaire
(GSEQ; Watson & Greenberg, 1996) consisted of
five 7-point Likert items: Three taken from Orlinsky
and Howard’s (1975) therapy-session evaluation
measure (inter-item reliability was .75), and two
items tapping a factor reflecting task helpfulness
(Elliott, 1985) in therapy (inter-item was reliability
.83; Warwar, 1995). These questions asked: (i) How
clients globally felt about the session that had
occurred, (ii) the degree to which they found their
therapist helpful, (iii) the degree of progress they felt
they were making as a result of the session, (iv) the
degree to which they experienced a change or shift as
the result of the session, and (v) the degree to which
they felt they wanted to take a new course of action
as a result of the session.
Client task-specific measure. The client task-
specific measure (CTSM) tracks post-session pro-
gress on the main tasks of treatment (Greenberg &
Safran, 1991; Greenberg et al., 1993). It consists of
12 items that are rated by clients on a 7-point Likert
scale. Three items each are related to: Self-critical
processes, problematic reactions, unfinished busi-
ness with a significant other, and experience of the
therapist’s client-centredness, respectively.
Outcome Measures
The beck depression inventory. This is a
21-item self-report inventory designed to measure
severity of depression (Beck, 1972; Beck et al.,
1961). Higher scores reflect greater severity of
depression (range = 0–63). Beck, Steer, and Garbin
(1988) report validity coefficients ranging from .66
to .86, and internal consistency coefficients ranging
from .73 to .93. Outcome scores for the BDI in this
study were clients’ residual gain scores calculated
from the larger combined York 1 and 2 samples.
Analyses were also duplicated using post-therapy
BDI scores.
Procedure
Defining alliance groups. The sample of 74
participants from the original two randomized con-
trolled trials was rank-sorted from LA to HA
session-one WAI scores. A WAI score of 5 or above,
indicating that clients are at a minimum “often”
agreeing with WAI items, was the criterion for a high
alliance (Pos & Thompson, 2010). Thirteen clients
had low alliances with mean WAI scores below 5,
and 18 clients had WAI scores of above 5. The WAI
scores in the HA group ranged from 5.42 to 7.00 (M
= 6.11, SD = .50), and WAI scores in the LA group
ranged from 3.58 to 4.67 (M = 4.19, SD = .37).
Scoring EP. EP ratings used in this study were
archival from Pos et al.’s (2009) study. EP was
measured as experiencing during EEs (EE-EXP; see
Pos et al., 2009). EEs (Greenberg & Korman, 1993)
are segments of psychotherapy in which clients speak
about having experienced emotion in response to a
real or imagined situation (see Greenberg & Korman,
1993; Pos et al., 2003). Reliability between raters for
sampling EEs was excellent. Pos et al. (2009) reported
that raters agreed 92% of the time on the identified EE
and its protocol. EP in the present study was defined
as average modal (most frequently expressed) EE-
EXP across all EEs from the working-phase sessions
from Pos et al.’s (2009) study. These were two
sessions between the fourth and fourth to last sessions
(on average between sessions 8 and 10) which clients
had reported as having been the most helpful. Help-
fulness was defined as that session having resulted in
the most progress, the highest degree of shift or
change (on the GSEQ), and the highest degree of
task resolution (on the CTSM). The average inter-
rater reliability reported for EE-EXP ratings was
excellent (weighted Cohen’s Kappa reported = .79;
Pos et al., 2009).
 Psychotherapy Research 7

MEE training. Six MEE raters (graduate stu-
dents and clinical faculty) received 40 hr of training
from MEE experts and Jeanne Watson, the scale
developer, from the University of Toronto. Inter-
class correlation coefficients between trained raters
and experts for ratings on 13 five-min video seg-
ments were used to establish reliability, and ranged
from .89 to .99 (excellent).
MEE rating procedure. As required by the
MEE manual, the 31 videos of first sessions were
each segmented into 5-min time bins for rating (i.e.,
1-hr video = 12 five-min time bins). For the present
study, the raters who were most reliable with the
MEE experts from the training phase independently
and fully rated all 5-min time bins within a session
video on each of the 10 scale dimensions. Disagree-
ments were discussed. Consensual ratings were
obtained and used in the analysis. Rater’s pre-
consensual ratings were used to establish reliability.
Downloaded by [Alberta Pos] at 09:12 09 May 2014
Therapist Expressed Empathy
Mean ratings and standard deviations of observer
ratings of therapist expressed empathy measured as a
total MEE score as well as by subscale scores are
presented in Table I. Consistent with expectations
driven by the experiential therapy model, the mean
global rating of therapist expressed empathy as well
as mean ratings for all 10 MEE dimensions across
therapists was in the high range (>6 or rated as
occurring over 75% of the time). Therapist dimen-
sions of “looking concerned” and “responsiveness”
had the highest average ratings and lowest standard
deviations, indicating that these therapist behaviours
were rated as the ones most consistently present.
The dimensions of “matching” and “attunement” had
the lowest average ratings and highest standard devia-
tions, indicating that experiential therapists varied
most on providing these empathic dimensions.

Hypothesis One: Session one Therapists’

Expressed Empathy Will Significantly Relate
Coding procedure. In order to arrive at session-
to Clients’ Post-Session Alliance Reports
level empathy ratings for the MEE dimensions, the
ratings for all time bins for each dimension across Hypothesis one was supported. An independent
the session were averaged. A total empathy score was samples t-test indicated significant differences in
then calculated as an average of the dimension therapist expressed empathy between the high (HA)
scores. versus low alliance (LA) groups, t(28) = 3.15,
p < .01. Therapists in the LA group were rated
significantly lower on average total expressed
Results empathy (M = 6.44, SD = .53) than those in the
HA group (M = 6.93, SD = .32; Table II)
The Measure of Expressed Empathy
The relationships among ratings on both total
Inter-rater reliability. Inter-rater agreement for MEE and dimensions of the MEE during session
total MEE scores was excellent, intraclass correla- one and clients’ post-session one alliance reports can
tion coefficient (ICC) = .85, and ranged from good be found in Table III. Due to the exploratory nature
to excellent for each of the dimensions, ICCs = of this study, the Pearson correlations were not
.62–.82 (Portney & Watkins, 2000; Shrouf & corrected for family-wise error and should be inter-
Fleiss, 1979). preted accordingly. Higher session-one therapist
global MEE scores related to clients reporting more
positive therapeutic alliances with their therapist by
Internal consistency. Cronbach’s alphas for
each of the 10 subscales and total MEE scores were
Table I. Subscale and total scale means of the MEE.
calculated. Internal consistency for the MEE total
score was high, α = .98, and ranged from good to MEE dimension Mean SD Max Min
excellent (α = .72 to α = .99) for the individual
MEE total 6.74 0.48 7.00 3.58
dimensions.
Vocal concern 6.73 0.46 7.48 5.00
Vocal expressiveness 6.62 0.48 7.50 5.36
Factors. Two factor analyses were conducted to Matching 6.58 0.59 7.48 4.75
Warmth/safety 6.68 0.55 7.58 5.03
confirm that the MEE constructs were loaded onto a
Attunement 6.59 0.68 7.53 4.61
single factor (empathy). One employed the max- Looking concerned 6.93 0.33 7.43 5.78
imum likelihood method and the other employed a Responsiveness 6.82 0.40 7.46 5.67
principle components analysis. Neither analysis Understanding experience 6.82 0.54 7.48 4.72
Understanding meaning 6.80 0.56 7.60 5.11
yielded significant sub-factors amongst the 10 dimen-
Genuineness/valuing 6.81 0.52 7.63 5.56
sions, indicating one overall expressed empathy
factor. Note. N = 30.
 8 A. J. Malin and A. E. Pos
Table II. Therapist ratings on the MEE as a function of WAI with clients’ WPEP. Therapist “vocal expressive-
group. ness” and “understanding experience” had the lar-
Alliance group Mean MEE SD Max Min gest relationships with WPEP.
A hierarchical regression analysis determined
LA group (mean WAI = 4.19, 6.44 0.53 7.32 5.16 whether client-rated alliance after session one could
SD = 0.37) predict clients’ WPEP once the early alliance was
HA group (mean WAI= 6.11, 6.93 0.32 7.37 6.31
SD = 0.50)
controlled. These results are presented in Table IV.
In the full final model, neither therapist expressed
Note. N = 30. empathy nor client-rated session-one alliance inde-
LA, low alliance; HA, high alliance. pendently predicted clients’ WPEP. The overall
model containing both variables was not significant,
the end of that session (r(28) = .54, p < .01). Higher F(2, 27) = 2.7, p = .11. There was a trend (p = .07)
therapist scores on each of the 10 MEE dimensions however, for therapists’ MEE scores predicting
were also significantly and moderately positively clients’ WPEP scores.
correlated with post-session one WAI scores. There- Another hierarchical regression tested the possib-
fore, higher ratings on all 10 therapist behavioural ility that the two empathic dimensions with the
dimensions related to significantly higher client- strongest relationships with WPEP might indepen-
rated post-session one alliances. Therapist matching, dently predict WPEP once the early alliance was
warmth/safety, and attunement had the largest cor- controlled. A combined score for “vocal expressive-
relations with the WAI scores, while therapists’ look ness” and “understanding experience” was used in
of concern had the smallest correlation with the WAI this analysis. Results are presented in Table V. The
Downloaded by [Alberta Pos] at 09:12 09 May 2014

scores. overall model was significant, F(2, 27) = 4.37,

Hypothesis Two: Therapists’ Expressed
Empathy in Session One Will Relate to Clients’
Working Phase Emotional Processing
Hypothesis two was also supported. Pearson correla-
tions between therapists’ session-one expressed
empathy and clients’ WPEP are also presented in
Table III. Therapists rated by observers as higher on
total expressed empathy in session one were more
likely to have clients with deeper modal EE-EXP
during the working phase of therapy. All but two
dimensions (“responsiveness” and “understanding
meaning”) showed significant positive relationships
p < .05, and the combined score of these two MEE
dimensions did significantly and independently pre-
dict clients’ WPEP, accounting for 24% of the
variance. Clients’ post-session one WAI ratings did
not significantly predict WPEP in this regression.
Hypothesis Three: Therapist’s Expressed
Empathy in Session One will Directly Predict
Outcome
Hypothesis three was not supported. Pearson corre-
lations of the relationships between therapists’
expressed empathy in session one, session-one alli-
ance, WPEP, and outcome on the BDI are presented

Table III. Pearson correlations among session-one expressed empathy, post-session one working alliance and WPEP.

Process variables 2 3 4 5 6 7 8 9 10 11 12 13

(1) WPEP .44* .44* .39* .49* .39* .42* .39* .39* .35 .46* .36 .40*
(2) WAI .54** .45* .46* .55** .55** .54** .40* .49* .51** .52** .46*
(3) MEE total .92 .92 .98 .99 .96 .84 .77 .97 .95 .93 .92
(4) Vocal concern – .89 .89 .91 .80 .85 .64 .91 .81 .86 –
(5) Vocal expressiveness – – .91 .94 .83 .80 .67 .86 .78 .85 –
(6) Matching – – – .98 .95 .77 .68 .93 .92 .92 –
(7) Warmth/safety – – – – .94 .82 .73 .94 .91 .93 –
(8) Attunement – – – – – .69 .72 .93 .97 .90 –
(9) Looking concerned – – – – – – .66 .82 .75 .73 –
(10) Responsiveness – – – – – – – .78 .72 .54 –
(11) Und. experience – – – – – – – – .94 .86 –
(12) Und. meaning – – – – – – – – – .88 –
(13) Genuine/valuing – – – – – – – – – – –

Note. N = 30. WPEPm = average modal experiencing during emotion episodes during working phase sessions; working phase = sessions
identified by clients as the most helpful on session measures; MEE dimensions: Vocal concern, vocal expressiveness. Matching, warmth/
safety, attunement, looking concerned, responsiveness. Und. experience, und. meaning, genuine/valuing.
WAI, Working Alliance Inventory; MEE, measure of expressed empathy.
*p < .05; **p < .01. When not corrected for family-wise error, simple Pearson r > .39 is significant at p = .05, r > .54 is significant at p = .01.

Table IV. Hierarchical regression predicting WPEP with therapist
expressed empathy and the working alliance.
Variables Total R2 ▵R2 F change df
Step 1 .19 .19 6.52* 1, 28
Therapist expressed
empathy
Step 2 .19 .00 .19 1, 27
Therapist expressed
empathy
Working alliance S1
Note. Therapist expressed empathy, measured by MEE. Working
alliance formation, measured by WAI. WPEP, measured by EXP
scale.
*p < .05.
in Table VI. Both post-session one WAI and work-
ing phase EE-EXP scores were significantly and
positively correlated with BDI residual gain scores,
while therapists’ MEE scores were not. Therefore,
no direct relationship between therapist expressed
empathy in session one and outcome for depressive
Downloaded by [Alberta Pos] at 09:12 09 May 2014
symptoms was indicated.
Indirect prediction. Since expressed empathy
was directly related to client-rated alliances and
clients’ WPEP, and both of these processes directly
predicted outcome, one remaining question was
whether the alliance and EP variables mediated
significant indirect effects of expressed empathy on
outcome. These concurrent indirect effects were
explored using path analysis using AMOS (Arbuckle,
2007) and Sobel tests. Therapist expressed empathy
was found to significantly affect outcome through
two indirect effects: One on post-session one WAI
ratings, Sobel z = −2.14, p = .02, and the other on
WPEP scores, Sobel z = −1.79, p = .04. The final
path diagram of processes predicting outcome is
shown in Figure 1. Given the sample size, we ran
bootstrap analyses as suggested by Preacher and
Hayes (2008) which yielded similar results (esti-
mates remained unaffected).
The hypothesized, AMOS model also had good fit
statistics (see Table VII for fit statistics with accept-
able values). Therefore, high therapist expressed
Table V. Hierarchical regression predicting WPEP with two dimensions of therapist expressed empathy and the working alliance.
Variables
Step 1
Vocal expressiveness + Understanding experience
Step 2
Vocal expressiveness + Understanding experience
Working alliance S1
Note. Vocal expressiveness + understanding experience, combined MEE variable. Working alliance formation, measured by the WAI.
Working phase emotional processing, measured by EXP scale.
*p < .05; **p < .01.
Psychotherapy Research 9
empathy in session one significantly contributed to
more favourable psychotherapy outcomes by con-
β tributing to two other therapy processes.
.44* Discussion
Experiential therapists are mandated to provide
.39 empathic attunement from the first moments of
.09
therapy in order to facilitate both positive working
alliances and clients’ EP. This study is the first to
test experiential assumptions concerning empathy’s
role in strengthening the early therapy alliance,
clients’ EP, and outcomes in experiential treatment
for depression.
Therapist Expressed Empathy and Early
Alliance Formation
Consistent with experiential theory and as originally
hypothesized, therapist empathy expressed during
the first session as rated by observers directly
predicted clients’ post-session one alliance reports.
Clients who had therapists rated as more empathic in
first sessions tended to rate their working alliance
with their therapist more favourably. This was true
for specific empathic dimensions as well as global
empathy operationalized as an average of 10
empathic dimensions. “Vocal matching,” “therapist
warmth and interpersonal safety,” and “responsive
attunement” are the empathic behaviours that were
most positively associated with session one alliance
formation. These empathic behaviours are also
identified in infant and neuropsychological litera-
tures as important to the development of emotional
regulation capacities (Porges, 2004; Schore, 2001;
Stern, 2000). Therefore, it may be that these ther-
apist behaviours in particular help clients emotion-
ally regulate from the first moments of therapy, and
may also be especially important to quickly building
a therapeutic relationship. Since the alliance as
measured by the WAI is operationalized as feeling
bonded with the therapist, and also as agreeing on
tasks and goals of therapy, a question might be
whether empathy contributed especially to one of
Total R2 ▵ R2 F change df β
.24 .24 8.90* 1, 28
.49**
.24 .00 .12 1, 27
.46*
.07
 10 A. J. Malin and A. E. Pos
Table VI. Pearson correlation matrix relating therapist expressed
empathy, the working alliance, WPEP and outcome.
Variables 1 2 3 4
MEE total – .54** .44* −.26
WAI – – .30 −.52**
WPEPm – – – −.49**
BDIres – – – –
MEE, measure of expressed empathy; WAI, working alliance
inventory; WPEPm, modal working phase EE-EXP score; BDIres,
BDI residual gain score.
*p < .05; **p < .01.
these subscales of the WAI; however, expressed
empathy did not relate more strongly to any subscale
in the present study. This may be the result of how
related the experience of a safe bond and agreement
on EP as both task and goal of therapy are in an
experiential approach. In another model of therapy
(dynamic therapy perhaps), therapist expressed
empathy might relate more strongly to the experi-
Downloaded by [Alberta Pos] at 09:12 09 May 2014
ence of the therapy bond (Gelso & Carter, 1985).
Therapist Expressed Empathy and WPEP
This study also tested the experiential assumption
that empathy contributes to an environment in
which clients can turn their inward attention towards
processing their emotional experiences. Clients with
therapists rated as more empathic in the first session
were those who were processing their emotions
significantly better, approximately eight sessions
later. This is important given that WPEP has been
found to be the most important direct predictor of
outcome in experiential therapy for depression (Pos
.43**
MEE_S1 WAI_S1
–.37*
. 54***
.38
–.41**
BDI
WPEPm
Figure 1. Final model predicting BDI residual gain scores.
MEE_S1, therapist scores on the measure of expressed empathy
in session 1; WPEPm, modal working phase emotional processing
scores on the EXP scale; WAI_S1, clients’ post-session one
ratings on the working alliance inventory; BDI, residual gain
scores on the BDI. Significant indirect effect of MEE on BDI
outcome through post-session one WAI ratings, Sobel z = −2.14,
p = 0.02, and significant indirect effect of MEE on BDI outcome
through working phase emotional processing scores, Sobel z =
−1.79, p = .04. *p < .05; **p < .01; ***p < .001.
Table VII. Summary of goodness of fit indices for the path
analysis model.
Model p for χ2 RMSEA CFI SRMR
BDI .534 .00 1.00 .020
Note. Good fit indicated by a chi-square p value greater than .05, a
root-mean-square error of approximation (RMSEA) value less
than or equal to .05 (<.08 indicating reasonable fit), a comparative
fit index (CFI) greater than .90 (>.95 very good fit), and a
standardized root-mean-square residual (SRMR) less than .05
(<.08 for reasonable fit).
BDI, Beck Depression Inventory.
et al., 2009). How can this finding be explained?
First, it is possible that therapists who are more
empathic at the opening moments of therapy pro-
vided their clients with early scaffolding, or an EP
“head-start” which maintained itself into the working
phase of therapy. Alternatively, if some therapists
were generally more empathic than others (i.e., if
empathy is a therapist trait), those able to be
empathic in first sessions may also have been those
who were able to provide consistent empathic sup-
port throughout therapy, including during working
phase sessions. However, another possibility is that
clients with an early propensity to process emotions,
or those clients who are well suited to an experiential
model of treatment, may have been easiest for
therapists to empathize with from the beginning of
therapy. Future empirical research may clarify these
possibilities.
Also interesting was the finding that not all
empathic behaviours captured by the MEE predicted
EP. First-session “responsiveness/following” and
“understanding meaning” were not associated with
WPEP, whereas therapist “vocal expressiveness” and
“understanding experience” amongst others were.
Therefore, while the sample size and the number of
correlations do not allow strong conclusions to be
made here, some limited evidence does suggest that
future research might further “parse” the MEEs
empathic dimensions to more precisely explore how
each MEE dimension contributes to other therapy
processes.
We also examined whether early therapist expressed
empathy predicted WPEP when the early alliance was
controlled, finding that a combined score for two
MEE subscales “vocal expressiveness” and “under-
standing experience” predicted WPEP over and above
the effect of session-one alliances. This highlights the
potential importance of these two empathic behaviours
for facilitating EP in clients.
Therapist Expressed Empathy and Outcome
The fact that no significant direct relationship
between first-session therapists expressed empathy

and outcome was found, but that post-session one
working alliances and WPEP directly predicted
outcome was consistent with past research by Pos
et al. (2009). Still, indirect effects of therapist expressed
empathy on outcome through the early alliance and
later EP variables were found. This indicates that the
expression of empathy by the therapist may in fact
fundamentally contribute to a number of other
important therapy processes. As such, therapist
empathy may be a necessary “ingredient” for other
therapy processes. Whether empathy is necessary for
strong alliances and EP across other therapy models
again will require further exploration.
The Measure of Expressed Empathy
This study also examined the validity and reliability
of the MEE (Watson & Prosser, 1999/2002), a
relatively new observer-rated measure. Results sup-
ported the MEE as a valid and reliable measure of a
Downloaded by [Alberta Pos] at 09:12 09 May 2014
coherent global construct of therapist expressed
empathy, with all 10 indicators of verbal and non-
verbal empathic behaviours reliably loading on a
global empathy construct. Of central interest is that
observer-rated measures often have not been as
successful as client-rated measures at capturing
essential processes. This has always been problem-
atic for both understanding and training in the
empathic process in psychotherapy. The fact that
empathy operationalized by the MEE directly pre-
dicted other important therapy processes (rated by
both clients and observer raters of emotional pro-
cess) and indirectly predicted outcome speaks well
for the conceptual validity of this observer measure.
As such, the MEE will have both important prac-
tical and empirical applications. Most helpful for
trainers of empathic process is that the MEE’s 10
dimensions outline specifically how therapists
express empathy. Thus, they have the potential to
be used by therapy supervisors as guides for provid-
ing concrete feedback to trainees concerning their
expression of empathy. A possible problem for the
MEE’s research application, however, may be its 10
dimensions which make video-coding labour intens-
ive. Future research could explore whether all 10
empathic dimensions are essential for addressing
research objectives. Just as some MEE subscales
here were found to correlate more strongly with EP
than others, other MEE scales may relate to pro-
cesses other than empathy per se. For example, one
could argue that a therapist cannot be empathic
unless he/she is present in the room perceiving the
client.
Psychotherapy Research 11
Therapist and Client Effects on the Expressed
Empathy Process
Several research groups have begun to explore
therapy processes in a manner that decomposes an
overall effect of any process into therapist effects
(differences between therapists’ abilities to engage in
a particular process independent of client differ-
ences) and client effects (differences among clients’
capacities to engage in a given process independent
of therapist differences; Baldwin, Wampold, & Imel,
2007; Del Re, Flückinger, Horvath, Symonds, &
Wampold, 2012). For example, if therapists see
more than one client, one can establish whether
their average performance on expressing empathy
results in explaining significant outcome variance
independent of the various clients each therapist
sees. On the other hand, some clients may have
difficulties being empathized with independent of
the therapist they see (within therapist variability in
expressed empathy independent of therapist differ-
ences). An interaction between client and therapist
effects is also possible. In this study, seven therapists
had two to six clients each, while nine had only one
client. To detect between (therapist effects) and
within therapist effects (client effects) in the present
sample would require estimating a substantial
amount of data for nine (56%) therapists for whom
therapist and client effects cannot be partitioned. We
did not have confidence that such analyses would
yield useful data. Moreover, even if differences in
average expressed empathy between therapists were
validly found, this would not guarantee that
expressed empathy emerges from a personality trait
of the therapist. Experiential therapy trainers would,
in fact, argue against this strong interpretation
having trained many therapists in empathic process
for decades (Goldman et al., 2006; Greenberg &
Watson, 2006; Watson et al., 1998). We would
argue that before exploring empathy as a therapist
effect, empathy must first be established as a ther-
apist “trait” variable. While within some models of
psychotherapy empathy may not be developed in
therapists, whether empathy is a trainable skill has
not, as yet, been empirically established. Here, we
have only done the foundational work of demon-
strating the importance of therapists expressing
empathy from the first moments of experiential
therapy.
That some clients may be easier to empathize with
than others, we feel, is an important consideration as
well. We did find, for example, that therapists who
had clients high in pre-therapy, non-assertiveness (as
measured by the Inventory of Interpersonal Pro-
blems; Horowitz, Rosenberg, Baer, Ureno, & Villa-
sernor, 1988) were rated by observers who were
 12 A. J. Malin and A. E. Pos
blind to clients’ interpersonal problems as being
significantly less empathic (Malin & Pos, 2012).
This suggests the possibility of a client effect, or at
minimum an interaction between therapist and client
effects. As such, it would not surprise us if expressed
empathy depends on a complex relationship between
the therapist and client, each of whom brings to
therapy personal characteristics and history (Baldwin
et al., 2007; Del Re et al., 2012).
We admit therefore that we do not here account
for the complex mechanisms by which expressed
empathy may arise. This suggests that future
research must explore the exact manner in which
differences in both therapist and client impact
therapist expressed empathy. In any event, the
present study points to the importance of therapists
expressing empathy from the first moments of
therapy to establishing early alliances, promoting
clients’ later EP, and to clients’ final outcomes.
In relation to testing therapist and client effects,
future research would have to employ larger samples
Downloaded by [Alberta Pos] at 09:12 09 May 2014
in which all or most therapists involved have mul-
tiple clients. Additionally, a concurrent important
avenue of future research will be to investigate
whether between therapist differences in expressed
empathy can be explained by differences in the
development of empathic skill, as well as explicate
what development or training in empathic ability will
involve.
Another consideration for future research is that
the present study examined expressed empathy as an
average therapist behaviour across first sessions.
Patterns of expression of empathy within sessions
were not examined. It is possible that a therapist may
have early difficulty empathizing with a particular
client and then recover, or wax and wane in their
provision of empathy within sessions. An important
future direction will be to examine the importance of
these patterns of expressed empathy unfolding
within and across sessions.
Finally, this study examined the relationships
between processes and outcome for experiential
therapy for depression only. As such, our results
may not generalize to other therapies or disorders.
However, since empathy has been articulated as a
common factor in all therapies (Norcross, 2011) and
since the MEE is comprised of behavioural corre-
lates of empathy found important in a variety of
different psychotherapies, one could argue that a
generally important change process applicable to
many approaches of therapy has been captured
here. Studies using the MEE in other therapy
models will be needed to confirm this.
Limitations. All studies have limitations. First,
although a sample of 30 clients is comparable to
samples typically used in process research (Critch-
field, Henry, Castonguay, & Borkovec, 2007; Henry,
Schacht, & Strupp, 1986, 1990; MacDonald, Cart-
wright, & Brown, 2007), small samples limit statist-
ical power and, as discussed earlier, our sample did
not allow for adequate exploration of therapist or
client effects. Still, power was not an issue for our
present research questions as significant results were
found, even when employing path analysis.
A second limitation is that the sample included
only clients with HA and LA scores. The study’s
conclusions may, therefore, not apply to clients with
average alliances. As a first study on observer-rated
expressed empathy, our main goal was testing
experiential theory concerning the importance of
empathy to alliance building and EP. A previous
study (Pos et al., 2009) with a larger sample (n = 74;
including the clients in this present study) has
already shown the session-one alliances and WPEP
predicted outcome in experiential therapy. That
expressed empathy predicted these two therapy
processes for these clients is in itself an important
result.
Third, some researchers and theorists may dislike
the fact that the MEE does not adequately assess
empathy as an interpersonal event, focusing more on
empathy as expressed by the therapist than empathy
viewed as an emergent between-person phenom-
enon. With the exception of the “therapist matching”
dimension, this appears to be true. The MEE does
not capture potential interpersonal empathic mar-
kers such as synchrony between the therapist and
client. Again it is an empirical question whether
interpersonal markers of empathy will be important
for our future understanding of empathy. Ramseyer
and Tschacher (2011) have shown that non-verbal
synchrony between therapist and client predicts
session alliances.
Conclusions
By using and further validating a new observer-rated
measure of therapist empathy (MEE; Watson &
Prosser, 1999/2002), the current study tested the
importance of therapists’ early expressions of
empathy to early alliance formation, later WPEP,
and positive experiential psychotherapy outcomes.
This study has provided empirical support for
theorized relationships in experiential theory
amongst these variables, as well as pointed to the
possibilities for future research to test the importance
of therapist expressed empathy as a common factor
in other models of treatment.

References
American Psychiatric Association (1987). Diagnostic and statistical
manual of mental disorders: DSM-III-R. Washington, DC:
Author.
Arbuckle, J. L. (2007). Amos (Version 16) [Structural equation
modeling software]. Chicago, IL: SPSS.
Baldwin, S. A., Wampold, B. E., & Imel, Z. E. (2007). Untan-
gling the alliance-outcome correlation: Exploring the relative
importance of therapist and patient variability in the alliance.
Journal of Consulting and Clinical Psychology, 75, 842–852.
doi:10.1037/0022-006X.75.6.842
Barkham, M., & Shapiro, D. A. (1986). Counsellor verbal
response modes and experienced empathy. Journal of Counseling
Psychology, 33(1), 3–10. doi:10.1037/0022-0167.33.1.3
Barrett-Lennard, G. T. (1962). Dimensions of therapist response
as causal factors in therapeutic change. Psychological Mono-
graphs, 76(43). (Whole No. 562.) doi:10.1037/h0093918
Barrett-Lennard, G. T. (1986). The relationship inventory
now: Issues and advances in theory, method and use. In
L. S. Greenberg & W. M. Pinsof (Eds.), The psychotherapeutic
process: A research handbook (pp. 439–476). New York, NY:
Guilford.
Barrington, B. L. (1961). Prediction from counsellor behavior of
client perception and of case outcome. Journal of Counseling
Downloaded by [Alberta Pos] at 09:12 09 May 2014
Psychology, 8(1), 37–42.
Beck, A. T. (1972). Depression: Causes and treatment. Philadelphia,
PA: University of Philadelphia Press.
Beck, A. T., Steer, R. A., & Carbin, M. G. (1988). Psychometric
properties of the beck depression inventory: Twenty-five years
of evaluation. Clinical Psychology Review, 8(1), 77–100.
doi:10.1016/0272-7358(88)90050-5
Beck, A. T., Ward, C. H., Mendelsohn, M., Mock, J., & Erbaugh,
J. (1961). An inventory for measuring depression. Archives of
General Psychiatry, 4, 561–571. doi:10.1001/archpsyc.1961.
01710120031004
Bohart, A. C., Elliott, R., Greenberg, L. S., & Watson, J. C.
(2002). Empathy. In J. C. Norcross (Ed.), Psychotherapy
relationships that work: Therapist contributions and responsiveness
to patients (pp. 89–108). New York, NY: Oxford University
Press.
Bohart, A. C., & Greenberg, L. S. (1997). Empathy and
psychotherapy: An introductory overview. In A. C. Bohart &
L. S. Greenberg (Eds.), Empathy reconsidered: New directions in
psychotherapy (pp. 3–31). Washington, DC: American Psycho-
logical Association.
Caracena, P. F., & Vicory, J. R. (1969). Correlates of pheno-
menological and judged empathy. Journal of Counseling Psycho-
logy, 16, 510–515. doi:10.1037/h0028437
Carkhuff, R. R., & Berenson, B. G. (1967). Beyond counseling and
therapy. New York, NY: Holt Rinehart & Winston.
Critchfield, K. L., Henry, W. P., Castonguay, L. G., & Borkovec,
T. D. (2007). Interpersonal process and outcome in variants of
cognitive–behavioral psychotherapy. Journal of Clinical Psycho-
logy, 63(1), 31–51. doi:10.1002/jclp.20329
D’Augelli, A. R. (1974). Nonverbal behavior of helpers in initial
helping interactions. Journal of Counseling Psychology, 21,
360–363.
Del Re, A. C., Flückinger, C., Horvath, A. O., Symonds, D., &
Wampold, B. E. (2012). Therapist effects in the therapeutic
alliance-outcome relationship: A restricted maximum likeli-
hood meta-analysis. Clinical Psychology Review, 32, 642–649.
doi:10.1016/j.cpr.2012.07.002
Elkin, I., Shea, T., Watkins, J. T., Imber, S. D., Sotsky, S. M.,
Collins, J. F., … Parloff, M. B. (1989). National institute of
mental health treatment of depression collaborative research
Psychotherapy Research 13
program. Archives of General Psychiatry, 46, 971–982. doi:10.
1001/archpsyc.1989.01810110013002
Elliott, R. (1985). Helpful and nonhelpful events in brief
counseling interviews. Journal of Counseling Psychology, 32,
307–322.
Elliott, R., Bohart, A. C., Watson, J. C., & Greenberg, L. S.
(2011). Empathy. Psychotherapy, 48 (1), 43–49. doi:10.1037/
a0022187
Elliott, R., & Greenberg, L. S. (2007). The essence of process-
experiential/emotion-focused therapy. American Journal of Psy-
chotherapy, 61(3), 241–254.
Ferrari, A. J., Charlson, F. J., Norman, R. E., Patten, S. B.,
Freedman, G., Murray, C. J. L., … Whiteford, H. A. (2013).
Burden of depressive disorders by country, sex, age and year:
Findings from the global burden of disease study 2010.
PLos Medicine, 10(11), e1001547. doi:10.1371/journal.pmed.
1001547
Gelso, C. J., & Carter, J. A. (1985). The relationship in counseling
and psychotherapy: Components, consequences, and theoret-
ical antecedents. The Counselling Psychologist, 13, 155–243.
doi:10.1177/0011000085132001
Gendlin, E. (1997). Experiencing and the creation of meaning:
A philosophical and psychological approach to the subjective.
Evanston, IL: Northwestern University Press.
Goldfried, M. (2004). Integrating integratively oriented brief
psychotherapy. Journal of Psychotherapy Integration, 14(1), 93–
105. doi:10.1037/1053-0479.14.1.93
Goldman, R., Greenberg, L. S., & Angus, L. (2006). The effects
of adding specific emotion-focused interventions to the thera-
peutic relationship in the treatment of depression. Psychother-
apy Research, 16, 537–549. doi:10.1080/10503300600589456
Goleman, D. (1995). Emotional intelligence. New York, NY:
Bantam Books.
Greenberg, L. S. (1986). Change process research. Journal of
Consulting and Clinical Psychology, 54(1), 4–9. doi:10.1037/
0022-006X.54.1.4
Greenberg, L. S. (2002). Emotional-focused therapy: Coaching clients
to work through their feelings. Washington, DC: American
Psychological Association.
Greenberg, L. S., & Elliott, R. (1997). Varieties of empathic
responding. In A. Bohart & L. S. Greenberg (Eds.), Empathy
reconsidered: New directions in theory, research and practice
(pp. 167–186). Washington, DC: American Psychological
Association.
Greenberg, L. S., & Korman, L. (1993). Assimilating emotion
into psychotherapy integration. Journal of Psychotherapy Integ-
ration, 3(3), 249–265.
Greenberg, L. S., & Pascual-Leone, A. (2006). Emotion in
psychotherapy: A practice-friendly research review. Journal of
Clinical Psychology, 62, 611–630. doi:10.1002/jclp.20252
Greenberg, L. S., Rice, L. N., & Elliott, R. K. (1993). Facilitating
emotional change: The moment-by-moment process. New York,
NY: Guilford Press.
Greenberg, L. S., & Safran, J. D. (1991). Emotion, psychotherapy,
and change. New York, NY: Guildford.
Greenberg, L. S., & Watson, J. C. (1998). Experiential therapy of
depression: Differential effects of client-centered relationship
conditions and process-experiential interventions. Psychother-
apy Research, 8, 210–224.
Greenberg, L. S., & Watson, J. C. (2006). Emotion-focused therapy for
depression. Washington, DC: American Psychological Association.
Henry, W. P., Schacht, T. E., & Strupp, H. H. (1986). Structural
analysis of social behavior: Application to a study of interper-
sonal process in differential psychotherapeutic outcome.
Journal of Consulting and Clinical Psychology, 54(1), 27–31.
doi:10.1037/0022-006X.54.1.27
 14 A. J. Malin and A. E. Pos
Henry, W. P., Schacht, T. E., & Strupp, H. H. (1990). Client and
therapist introject, interpersonal process, and differential psy-
chotherapy outcome. Journal of Consulting and Clinical Psycho-
logy, 58, 768–774. doi:10.1037/0022-006X.58.6.768
Horowitz, L. M., Rosenberg, S. E., Baer, B. A., Ureno, G., &
Villasernor, V. S. (1988). Inventory of interpersonal problems:
Psychometric properties and clinical application. Journal of
Consulting and Clinical Psychology, 56, 885–892. doi:10.1037/
0022-006X.56.6.885
Horvath, A. O. (1981). An exploratory study of the working alliance
its measurement and relationship to therapy outcome (Doctoral
dissertation). Retrieved from ProQuest Dissertations and
Theses Database. NK55037.
Horvath, A., & Bedi, R. P. (2002). The alliance. In J. C. Norcross
(Ed.), Psychotherapy relationships that work: Therapist contribu-
tions and responsiveness to patients (pp. 37–69). London: Oxford
University Press.
Horvath, A., & Greenberg, L. S. (1986). The development of the
working alliance inventory. In W. M. Pinsof & L. S. Greenberg
(Eds.), Guilford clinical psychology and psychotherapy series
(pp. 529–556). New York, NY: Guilford Press.
Horvath, A., & Greenberg, L. S. (1989). Development and
validation of the working alliance inventory. Journal of Counsel-
ing Psychology, 36, 223–233. doi:10.1037/0022-0167.36.2.223
Klein, M. H., Mathieu-Coughlan, P., & Kiesler, D. J. (1986). The
Downloaded by [Alberta Pos] at 09:12 09 May 2014
experiencing scales. In L. Greenberg & W. Pinsof (Eds.), The
psychotherapeutic process (pp. 21–71). New York, NY: Guilford.
Lambert, M. J. (1992). Psychotherapy outcome research: Implica-
tions for integrative and eclectical therapists. In J. C. Norcross
& M. R. Goldfried (Eds.), Handbook of psychotherapy integration
(pp. 94–129). New York, NY: Basic Books.
Lambert, M. J., De Julio, S. S., & Stein, D. M. (1978). Therapist
interpersonal skills: Process, outcome, methodological consid-
erations, and recommendations for future research. Psycholo-
gical Bulletin, 85, 467–489. doi:10.1037/0033-2909.85.3.467
Macdonald, J., Cartwright, A., & Brown, G. P. (2007). A
quantitative and qualitative exploration of client-therapist
interaction and engagement in treatment in an alcohol service.
Psychology and Psychotherapy: Theory, Research, and Practice,
80, 247–268. doi:10.1348/147608306X156553
Malin, A. J., & Pos, A. (2012, June). The influence of clients’
interpersonal problems on their capacity to utilize therapist empathy
during experiential treatment of depression. Presentation at the
Society for Psychotherapy Research (SPR) Annual Conference,
Virginia Beach, VA.
Mosely, D. (1983). The therapeutic relationship and its association
with outcome (Unpublished master’s thesis). University of
British Columbia, Vancouver.
Myers, S. (2000). Empathic listening: Reports on the experience
of being heard. Journal of Humanistic Psychology, 40, 148–173.
doi:10.1177/0022167800402004
Norcross, J. (2011). Psychotherapy relationships that work: Evidence-
based responsiveness. New York, NY: Oxford University Press.
Orlinsky, D. E., & Howard, K. I. (1975). Varieties of psychother-
apeutic experience. New York, NY: Teacher’s College Press.
Paivio, S. C., & Laurent, C. (2001). Empathy and emotion
regulation: Reprocessing memories of childhood abuse. JCLP/
In Session: Psychotherapy in Practice, 57, 213–226.
Pascual-Leone, A., & Greenberg, L. S. (2007). Emotional
processing in experiential therapy: Why “the only way out is
through”. Journal of Consulting and Clinical Psychology, 75(6),
875–887. doi:10.1037/0022-006X.75.6.875
Porges, S. W. (2004). Neuroception: A subconscious system for
detecting threats and safety. Zero to Three, 24(5), 19–24.
Portney, L. G., & Watkins, M. P. (2000). Foundations of clinical
research: Applications to practice (pp. 560–567). Upper Saddle
River, NJ: Prentice Hall.
Pos, A. E. (2006). Experiential treatment for depression: A test of the
experiential theory of change, differential effectiveness, and predictors
of maintenance of gains (Doctoral dissertation). Retrieved from
ProQuest Dissertation and Theses Database. AAINR19820.
Pos, A. E., & Greenberg, L. S. (2007). Emotion-focused therapy:
The transforming power of affect. Journal of Contemporary
Psychotherapy, 37, 25–31.
Pos, A. E., Greenberg, L. S., & Elliott, R. (2008). Experiential
therapy. In J. Lebow (Ed.), Twenty-first century psychotherapies
(pp. 80–122). New York, NY: Wiley.
Pos, A. E., Greenberg, L. S., & Goldman, R. N. (2005). Depth of
emotional experience and outcome. Psychotherapy Research,
15, 248–260. doi:10.1080/10503300512331385188
Pos, A. E., Greenberg, L. S., Goldman, R. N., & Korman, L. M.
(2003). Emotional processing during experiential treatment of
depression. Journal of Consulting and Clinical Psychology, 71,
1007–1016. doi:10.1037/0022-006X.71.6.1007
Pos, A. E., Greenberg, L.S., & Warwar, S. H. (2009). Testing a
model of change in the experiential treatment of depression.
Journal of Consulting and Clinical Psychology, 77, 1055–1066.
doi:10.1037/a0017059
Pos, A. E., & Thompson, M. (2010). Exploring patterns of alliance
development during experiential therapy of depression: A comparison
of methods. Panel Presentation at the annual meeting of the
Society for Psychotherapy Research, Asilomar, CA.
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resam-
pling strategies for assessing and comparing indirect effects in
multiple mediator models. Behavior Research Methods, 40, 879–
891. doi:10.3758/BRM.40.3.879
Ramseyer, F., & Tschacher, W. (2011). Nonverbal synchrony in
psychotherapy: Coordinated body-movement reflects relation-
ship quality and outcome. Journal of Consulting and Clinical
Psychology, 79, 284–295. doi:10.1037/a0023419
Rice, L., Greenberg, L., & Watson, J. (1994). Manual for client-
centered therapy (Unpublished manuscript). Toronto, ON: York
University.
Rogers, C. R. (1957). The necessary and sufficient conditions of
therapeutic personality change. Journal of Consulting Psychology,
21(2), 95–103. doi:10.1037/h0045357
Rogers, C. R. (1959). A theory of therapy, personality, and
interpersonal relationships, as developed by the client-centered
framework. In S. Koch (Ed.), Psychology a study of a science
(Vol. 3, pp. 184–256). New York, NY: McGraw Hill.
Rogers, C. R. (1975). Empathic: An unappreciated way of being.
Counseling Psychologist, 5(2), 2–10. doi:10.1177/
001100007500500202
Salvio, M., Beutler, L. E., Wood, J. M., & Engle, D. (1992). The
strength of the therapeutic alliance in three treatments for
depression. Psychotherapy Research, 2(1), 31–36. doi:10.1080/
10503309212331333578
Schore, A. N. (2001). Effects of secure attachment relationship
on right brain development, affect regulation, and infant
mental health. Infant Mental Health Journal, 22 (1–2), 7–66.
doi:10.1002/1097-0355(200101/04)22:1%3C7::AID-IMHJ2%
3E3.0.CO;2-N
Shea, M. T., & Elkin, I. (1996). The NIMH treatment of depression
collaborative research program. London: Gaskell/Royal College of
Psychiatrists.
Shrouf, P. E., & Fleiss, J. L. (1979). Intraclass correlations:
Uses in assessing rater reliability. Psychological Bulletin, 86(2),
420–428. doi:10.1037/0033-2909.86.2.420
Spitzer, R., Williams, J., Gibon, M., & First, M. (1989). Structured
clinical interview for the DSM-III-R. New York, NY: American
Psychiatric Press.
Steckley, P. L. (2006). An examination of the relationship between
clients’ attachment experiences, their internal working models of self
and others, and therapists’ empathy in the outcome of process-

experiential and cognitive-behavioural therapies (Doctoral disser-
tation). Retrieved from ProQuest Dissertations and Theses
Database. NR15730
Stern, D. N. (2000). The interpersonal world of the infant. New York,
NY: Basic Books.
Teasdale, J. D. (1999). Emotional processing, three modes of
mind and the prevention of relapse in depression. Behaviour
Research and Therapy, 37, S53–S77.
Tepper, D. T. (1973). The communication of counselor empathy,
respect and genuineness through verbal and non-verbal channels
(Doctoral dissertation). University of Massachusetts, Amherst.
Tracey, T. J., and Kokotovic, A. M. (1989). Factor structure
of the working alliance inventory. Psychological Assessment,
1, 207–210. doi:10.1037/1040-3590.1.3.207
Truax, C. B., & Carkhuff, R. R. (1967). Toward effective counseling
and psychotherapy. Hawthorne, NY: Aldine Publishing.
Van Geert, P.L.C. (1996). The development of attachment and
attachment-related competences. A dynamic model. In: W.
Koops, J. Hoelsma, & J. van den Boom (Eds.), New develop-
ments in attachment research (pp. 181–199). Amsterdam, North
Holland: Royal Academy of Sciences.
Wampold, B. E. (2001). The great psychotherapy debate: Models,
methods, and findings. Mahwah, NJ: Lawrence Erlbaum
Associates.
Warwar, S. H. (1995). The relationship between level of experiencing
Downloaded by [Alberta Pos] at 09:12 09 May 2014
and session outcome (Unpublished master’s thesis). York Uni-
versity, Toronto, Ontario.
Watson, J. C. (2001). Re-visioning empathy. In D. Cain &
J. Seeman (Eds.), Humanistic psychotherapies: Handbook of
research and practice (pp. 445–471). Washington: APA.
Watson, J. C., & Greenberg, L. S. (1996). Pathways to change in
the psychotherapy of depression: Relating process to session
change and outcome. Psychotherapy: Theory, Research, Practice
and Training, 33(2), 262–274.
Watson, J. C., Enright, C., & Kalogerakos, F. (1998, June). The
impact of therapist variables in facilitating change. Paper presented
View publication stats
Psychotherapy Research 15
at the annual meeting of the Society for Psychotherapy
Research, Snowbird, UT.
Watson, J. C., Gordon, L. B., Stermac, L., Kalogerakos, F., &
Steckley, P. (2003). Comparing the effectiveness of process-
experiential with cognitive-behavioral psychotherapy in the
treatment of depression. Journal of Consulting and Clinical
Psychology, 71, 773–781. doi:10.1037/0022-006X.71.4.773
Watson, J. C., Goldman, R., & Vanaerschot, G. (1998).
Empathic: A post-modern way of being? In L. S. Greenberg,
J. C. Watson, & G. Lietaer (Eds.), Handbook of experiential
psychotherapy (pp. 61–81). New York, NY: Guilford Press.
Watson, J. C, & Prosser, M. (1999/2002). Measure of expressed
empathy (Unpublished manuscript). OISE, University of
Toronto, Toronto, ON.
Westen, D., & Morrison, K. (2001). A multidimensional meta-
analysis of treatments for depression, panic, and generalized
anxiety disorder: An empirical examination of the status of
empirically supported therapies. Journal of Consulting and
Clinical Psychology, 69, 875–899. doi:10.1037/0022-006X.69.
6.875
Wing, E. H. (2010). The relationship between therapist empathy, the
working alliance, and therapy outcome: A test of a partial mediation
model (Doctoral dissertation). Retrieved from ProQuest Dis-
sertation and Theses Database. AAI3398120.
Wong, K., & Pos, A. E. (2014). Interpersonal processes affecting
early alliance formation in experiential therapy for depression.
Journal of Psychotherapy Research, 24(1), 1–11. doi:10.1080/
10503307.2012.708794
World Health Organization. (2004). The burden of disease: 2004
update. Retrieved January 25, 2011, from the World Health
Organization http://www.who.int/healthinfo/global_burden_
disease/GBD_report_2004update_full.pdf.