Available online at www.sciencedirect.

com

Behavior Therapy 43 (2012) 36 – 48

www.elsevier.com/locate/bt

Common Factors of Change in Couple Therapy

Sean D. Davis
Alliant International University

Jay L. Lebow
Family Institute at Northwestern University

Douglas H. Sprenkle
Purdue University

cate that these variables account for far more

Though it is clear from meta-analytic research that couple
therapy works well, it is less clear how couple therapy
works. Efforts to attribute change to the unique ingredients
of a particular model have consistently turned up short,
leading many researchers to suggest that change is due to
common factors that run through different treatment
approaches and settings. The purpose of this article is to
provide an empirically based case for several common
factors in couple therapy, and discuss clinical, training, and
research implications for a common factors couple therapy
paradigm. Critical distinctions between model-driven and
common factors paradigms are also discussed, and a
moderate common factors approach is proposed as a more
useful alternative to an extreme common factors approach.
Keywords: couple therapy; common factors; outcome research
AS DISCUSSED IN THE introduction to this special
section (Halford & Snyder, 2012-this issue), the
efficacy of couple therapy is well established. What
is less clear, however, is how couple therapy
works. Data suggest that successful couple therapy
consists of a complex interaction of numerous
pantheoretical variables, and initial findings indi-
Address correspondence to Sean D. Davis, Alliant International
University, Marital and Family Therapy, 2030 West El Camino Avenue,
Suite 200, Sacramento, CA 95833; e-mail: sdavis2@alliant.edu.
0005-7894/xx/xxx-xxx/$1.00/0
© 2011 Association for Behavioral and Cognitive Therapies. Published by
Elsevier Ltd. All rights reserved.
variance than the unique contributions of any
particular model (Sprenkle & Blow, 2004). A
more thorough understanding of these pantheore-
tical variables, generally referred to as “common
factors,” how and when they are activated and
how they are woven into empirically based,
model-driven treatment could greatly simplify
and refine couple therapy research and training,
and ultimately maximize the effectiveness of
couple therapy. While the data point to the
importance of common factors, much of the
relevant research in couple therapy has yet to be
done. Nevertheless, enough data exist to make a
strong case for the ideas presented in this article.
The purpose of this article is to outline the main
tenets of a common factors paradigm, provide an
overview of what is known about the common
factors, and to outline research, training, and
clinical implications. Interested readers can find a
more in-depth discussion of these issues in
Sprenkle, Davis, and Lebow (2009).
Common Factors and Model-Driven Change:
Two Paradigms of How Couples Change
The common factors paradigm stands as an
alternative to the model-driven change paradigm.
Though we believe there are significant differences
between the two paradigms, we also acknowledge
that polarizing two paradigms overemphasizes
differences and underemphasizes similarities. We
outline below what we see as the polarities of these
two paradigms, and discuss later our preferred
“moderate” common factors approach. The
 common factors of change in couple therapy
traditional model-driven paradigm assumes that the
primary explanation for change rests within the
unique elements and mechanisms of individual
treatment models. Conversely, the common factors
paradigm assumes that common mechanisms of
change cut across all effective psychotherapies, and
that models are the vehicles through which
common factors are potentiated (Sprenkle &
Blow, 2004).
Taken to the extreme, the model-driven para-
digm suggests that treatments are “dispensed” in
ways analogous to drugs and medical procedures.
In contrast, a common factors approach is more
contextual, emphasizing the interaction of com-
plex variables affecting treatment such as client
and therapist variables, alliance, and expectancy.
These latter factors are viewed as more important
than the unique contributions of specific models.
The model-driven change paradigm often deem-
phasizes the therapist's role in change, instead
focusing on the treatment dispensed. The com-
mon factors approach, on the other hand,
emphasizes that treatment models do not exist
in therapy outside of the therapist delivering
them, and therefore the qualities of the therapist
delivering the treatment are more important than
the treatment itself.
The model-driven paradigm is more therapist-
centric in that it emphasizes the importance of the
therapist performing therapy in a certain manner,
and the client falling into step with what the
therapist recommends. Conversely, the common
factors paradigm is more client-centric in that
models—regardless of whether they are therapist
or client centered—are adapted to a client's
unique needs and world views, and clients use
whatever is offered in therapy in unique and
idiosyncratic ways not necessarily predicted by
the model.
The model-driven change paradigm has always
been the more prominent of the two approaches.
Models are more exciting and interesting than the
factors that all good therapies possess. Perhaps it
is no surprise that most couple therapy training
programs emphasize models and rarely directly
address common factors. Today, most federal
research funding goes toward this paradigm, and
advocacy of “evidence-supported treatments” has
become commonplace. The common factors ap-
proach, while enjoying intuitive appeal among
many clinicians and a growing number of re-
searchers, has a long way to go before it enjoys a
similar seat at the cultural table as the model-
driven change paradigm. We hope that this article
will help bring the common factors paradigm into
the mainstream.
37
The State of the Research Surrounding
Common Factors
Randomized clinical trials (RCTs) are widely
considered the gold standard for establishing
treatment efficacy. In both individual and couple
therapy (though there are far fewer RCTs in couple
therapy) RCTs routinely demonstrate that the
model being tested outperforms treatment as usual
and wait-list control conditions. Claims of efficacy
solidify as RCTs accumulate for a particular model.
However, meta-analytic reviews of RCTs suggest
that when RCTs are compared to each other,
differences in treatments largely disappear, partic-
ularly when controlling for confounding variables
such as researcher allegiance (Wampold, 2001). In
other words, demonstrating efficacy relative to
treatment as usual and wait-list controls does not
equal efficacy relative to other models.
Meta-analytic research that calculates an effect size
by comparing diverse studies is sometimes criticized as
confounding samples, treatments, presenting prob-
lems, and outcome measures. Inclusion of methodo-
logically weak studies can further confound meta-
analytic results. Nevertheless, most meta-analytic
reviews show few if any differences between treat-
ments, and several show that this finding is as strong
when comparing only methodologically sound
studies as it is when including weak studies (Shadish
& Baldwin, 2002; Smith & Glass, 1977). While there
are several possible explanations of this finding (e.g.,
different models may use different yet equally potent
change mechanisms), we believe the common factors
hypothesis is the most consistent with existing data.
Even most studies that represent the ideal situation—
an RCT with at least two established models that
controls for confounding variables such as researcher
allegiance—fail to show significant differences
among treatments (Elkin et al., 1989), but showed
significant differences among such common factors
as therapists characteristics (Blatt, Sanislow, Zuroff,
& Pilkonis, 1996). Increasingly the debate is not over
whether but how much common factors influence
outcome (Chambless & Ollendick, 2001; Orlinsky,
Grawe, & Parks, 1994).
As research methods become more complex, so
may our understanding of the change process.
Therefore, it may be better to expand RCTs rather
than do away with them altogether. We believe that
an investigation of common factors and processes
could be woven into an RCT, thus allowing for a
more refined view of the change process. However,
although reliable differences in treatment models
may be discovered, we believe it is unlikely that one
model will be shown to be universally more
effective than others. We believe it is more likely
38 davis et al.
that research will discover certain change processes
and mechanisms that mediate change during certain
stages of successful therapy, but that many different
models can use specific methods to facilitate those
change mechanisms. In fact, several couple therapy
RCT studies that incorporate process research
suggest that similar processes such as softening of
emotional expressiveness and responsiveness in
emotionally focused therapy (EFT; Johnson &
Greenberg, 1985) and communication, behavior
change in target problems, and acceptance of target
problems in integrative and traditional behavioral
couple therapies (IBCT, TBCT, respectively; Doss,
Thum, Sevier, Atkins, & Christensen, 2005)
mediate change. Conceptual overlap between the
processes measured in both studies could lend
support to the common factors hypothesis that
similar processes are impacted by different models
using different interventions.
Nuanced findings mentioned above may provide
insight into which processes are at work at which
stage of therapy with which types of clients and
presenting problems. Such research could suggest
that the focus on processes or principles (e.g.,
facilitating softening) take precedence over a focus
on models. In other words, the model (or component
of the model) would be selected based on the degree
to which it facilitated the requisite process, and the
requisite process would be determined by client
feedback (Halford et al., 2012-this issue), the stage of
therapy, unique issues relevant to the couple's
presenting problem, and other variables discovered
as research progressed. This client-centered ap-
proach would still be empirically informed, but
much more flexible and adaptive (and therefore
presumably effective) than the dominant “hammer/
nail” approach in which a client is expected to
conform to the efficacious treatment rather than vice
versa (Blow, Sprenkle, & Davis, 2007).
This approach to couple therapy could be similar
to Prochaska's widely used model in substance
abuse treatment that focuses on selecting the type of
treatment approach based on the client's “stage of
change” (Prochaska & Norcross, 2009). A similar
integrative couple therapy approach has been
advocated by Snyder (1999). We also applaud
Christensen's (2010) similar efforts to develop a
treatment protocol that uses five general principles
to guide the selection of different methods from
effective couple therapy approaches throughout the
course of treatment. We believe process- and
principle-based approaches such as these will be a
promising major direction in future couple therapy
practice, research, and training.
Indirect research supports process- and principle-
based approaches. Adapting treatment based on client
feedback, which presupposes the flexibility discussed
above, has been shown to be a strong predictor of
change (Halford et al., 2012-this issue). Some evidence
suggests a better fit between certain clients and
approaches, with lower-socioeconomic status clients
often preferring more directive, skills-focused ap-
proaches than their higher-socioeconomic counter-
parts (Cline, Mejia, Coles, Klein, & Cline, 1984).
Matching the type and pacing of interventions with
client characteristics also seems to be important, as
emotionally reactive clients have been shown to prefer
inductive, eliciting approaches with high structure,
whereas less reactive clients may prefer less structure
(Beutler, Malik, & Alimohamed, 2004). Furthermore,
some dyadic approaches do a better job of meeting
certain unique challenges (e.g., substance abuse,
Powers, Vedel, & Emmelkamp, 2008; depression,
Bodenman et al., 2008; intimate partner violence,
Stith, Rosen, McCollum, & Thomsen, 2004) than
individual approaches, though little is known about
how or why.
Our moderate common factors approach suggests
that because most models can facilitate similar
couple processes, it is likely that most models will
find a use either throughout or at different points in
the therapeutic process with most clients. This could
explain why most tested couple therapy models are
effective, but none significantly more than another—
because they all do an adequate job of invoking the
right change mechanisms at the right times. It could
also explain why treatment effects for many
couples decrease significantly at 2-year follow-up
(Halford & Snyder, 2012-this issue)—because
treatment was not nuanced enough to adequately
meet their long-term needs. It could also explain
the wide variation in therapist effectiveness (Blatt
et al., 1996)—some therapists may simply be better
at judging what clients need (or at eliciting client
feedback about the same) at key times (Halford et
al., 2012-this issue), and have a broader reservoir
of approaches to help treatment progress accord-
ing to the client's needs. Future research is needed
to test these hypotheses.
Common Factors Shared by Individual and
Couple Therapy
Most of the direct empirical support for common
factors comes from the individual therapy literature.
While much of what makes a good individual
therapist likely also makes a good couple therapist,
it is possible that there are differences (Blow et al.,
2007). However, in most cases the limited direct
inquiry into common factors in couple therapy tem-
pers assertions linking common factors in individual
therapy to couple therapy. Research reviewed below is
for individual therapy unless otherwise indicated.
 common factors of change in couple therapy
client variables
We believe that clients are the most important
variable in therapy. Bohart and Tallman (2010)
propose that the reason a number of approaches
work equally well is because clients are adept at
taking whatever a therapist offers and tailoring it to
suit their needs. An excellent therapist can master-
fully deliver the most up-to-date treatment avail-
able, but if the client lacks awareness, is unwilling
to take responsibility for changing, or otherwise
does not engage in treatment, it will all be for
naught. Conversely, a highly motivated, proactive
client who focuses on what their treatment has to
offer rather than where it is lacking will likely shape
whatever is offered to their betterment.
In keeping with our moderate common factors
approach (discussed later), we believe that therapists
and models almost always do matter. Models
provide a critical framework for treatment, without
which many of the common factors may not be
potentiated. For example, client motivation may
suffer if the therapist, as informed by their model,
does not offer clients a more adaptive way of
experiencing their problem. However, the ability to
convince the clients that the model is credible may be
more important than the actual content of the model.
The ability to present the approach convincingly
would be the common factor. A poor therapist could
thwart a motivated couple, just as an excellent
therapist could engage a couple with low motivation
or present most models in such a way that fit with the
client's worldview. Conversely, a poor fit between
the model and the client's worldview can lead to
treatment dropout even if the clients are otherwise
engaged and the therapist is typically effective
(Johnson & Talitman, 1997). Couples in which one
partner is “just visiting” or divorce is being actively
discussed often do poorly. Nonetheless, given that
most therapists are at least somewhat effective and
most models likely have something to offer most
people, client characteristics remain the most potent
determinant of outcome.
Despite this, research thus far sheds remarkably
little light on client variables that affect couple
therapy outcome. Clinical trials research typically
treats client variation as a source of error rather
than an opportunity for discovery. Furthermore,
studies often group couples based on diagnosis,
ignoring individual characteristics that may be a
larger contributor to change. Most research on
client variables in couple therapy focuses on static
characteristics such as age, gender, sexual orienta-
tion, and race. Few studies focus on variables more
likely related to outcome such as motivation, locus
of control, and engagement in treatment (Blow
et al., 2007). For example, Johnson and Talitman
39
(1997) found that EFT worked best with couples
who believed the tasks of EFT were related to their
problem. Jacobson and Christensen (1996) found
that traditional behavioral marital therapy worked
best with couples with high commitment and low
“traditionality” and similar goals for the marriage.
Some client characteristics appear to have an
influence on outcome. Individual comorbid psy-
chopathology has been associated with poorer
outcomes in behavioral couple therapy treatments
of alcohol use disorder, though less so than
individual treatments for the same disorder
(McGrady, Epstein, Cook, Jensen, & Hildebrandt,
2009). Initial levels of relationship distress, sexual
dissatisfaction, and increased length of marriage
have been negatively related to outcome in integra-
tive and traditional behavioral marital therapy
(Atkins et al., 2005). As discussed earlier, some
problems respond better to couples than individual
treatments. While conjoint treatment is clearly
indicated for some problems, many conjoint
approaches could prove effective. In other words,
while some problems may benefit more from a
specialized class of approach, several treatments
that share core characteristics could likely be
developed for that class. For example, conduct-
disordered children and their families typically
respond best to approaches that are structured,
ecologically/systemically based, highly involved,
focused somewhat on the individual teenager, and
use assertive methods of treatment engagement.
However, there are at least four such empirically
validated approaches, none of which have been
shown to be differentially effective relative to the
others. While future research may shed light on
which presenting problems or client characteristics
benefit from a specialized focus as well as the
necessary elements of that focus, we believe it is
unlikely that there will ever be a problem or client
characteristic for which only one theoretical ap-
proach is indicated.
Client motivation and engagement is likely to be
a more complex issue when dealing with two people
with disparate levels of motivation using disparate
methods to work toward sometimes disparate
goals. Davis and Piercy (2007a, 2007b), in their
qualitative study, outline several characteristics
exhibited by clients as they moved through the
process of successful couple therapy, as well as
what the therapists did to nurture these character-
istics and counter negative characteristics. These
included humility (operationalized as a willingness
to take personal responsibility for their role in the
problems and a willingness to risk/be vulnerable),
commitment to each other and the process of
therapy, psychological and systemic awareness, and
40 davis et al.
hard work. Despite promising preliminary findings,
we have much to learn about how client charac-
teristics influence outcome in couple therapy, and
if and/or how therapists can influence these
characteristics.
client feedback
Given the importance of the client's perception of
therapy and the tendency of therapists to misjudge
the client's experience of therapy it follows that
formally eliciting feedback and making requisite
adjustments to the course of treatment is another
potential common factor (Halford et al., 2012-this
issue). Receiving formal client feedback about the
progress of therapy has been shown to improve
outcome regardless of whether that feedback is
delivered via the clients themselves (Anker, Duncan,
& Sparks, 2009) or a clinical supervisor systemati-
cally monitoring client progress (Lambert et al.,
2001). Though the strong benefits of providing
treatment progress feedback is more established in
the individual therapy literature, initial findings in
the couple therapy literature suggest a similar pattern
(Halford et al.). Anker et al. found that couples
whose therapist received direct feedback about
treatment progress “achieved nearly 4 times the
rate of clinically significant change, and maintained a
significant advantage on the primary measure at 6-
month follow-up while attaining a significantly
lower rate of separation or divorce” (p. 693).
therapist variables
Buried among the data of what is likely the most
methodologically sound psychotherapy outcome
study to date, the National Institute of Mental
Health Collaborative Depression Study (Elkin
et al., 1989) is a surprising finding: despite strong
efforts to control for therapist factors (e.g., training,
experience level, model allegiance), there was far
more variability in effectiveness across therapists
than across treatment approaches. In fact, the most
consistently effective therapist was a psychiatrist
who split her work between the antidepressant drug
clinical management and placebo clinical manage-
ment conditions. She did not participate in either
the cognitive-behavioral or interpersonal therapy
conditions; her “therapy” consisted of meeting
briefly with clients with the main goal of being
supportive (Blatt et al., 1996)!
Wampold (2001), in his major meta-analysis of
studies comparing bona fide treatments, found that
differences among therapists contribute more to
outcome variance (effect size of 0.60) than the
treatment models they used (at most an effect size of
0.20, but probably closer to zero). Furthermore,
Type 1 errors are significantly inflated when
therapist variables are ignored, which makes
differences among treatments appear larger than
they really are. Others (Blatt et al., 1996) report
similar findings. These are therapists in well-
controlled outcome studies; it is likely that there is
far more variance among therapists in everyday
clinical practice.
There is so much variance among therapists that
Blow et al. (2007) propose that it may be better to
talk about empirically validated therapists than
empirically validated therapy. Therapeutic models
come alive or die largely through the therapist. As
common factors work through models, models
work through therapists. What, then, are common
factors of effective therapists? Surprisingly, little is
known about this question. Many of the variables
we review below account for relatively small
amounts of the variance (Blow et al., 2007).
Static characteristics (e.g., age, gender, race)
account for little variance, providing hope that a
competent therapist can overcome whatever initial
limitations may be presented by such variables
(Blow et al., 2007). Experience level is another
variable that has surprisingly shown to not account
for much of the variance. In one study (Stolk &
Perlesz, 1990), first-year students performed better
than second-year students, possibly because first-
year students focused more on forming an alliance
and second-year students, having gotten more
comfortable in the room, focused more on tech-
nique at the expense of the alliance. It is also
possible that experienced therapists work with
more difficult cases.
A friendly, positive therapist tends to get better
results than a critical or hostile therapist, or a
therapist who is defensive, particularly early in
therapy. Effective therapists also tend to be sensitive
to the unique needs of their clients and flexible in
their treatment approaches, offering insight-oriented
interventions to more reflective clients, and skill-
building interventions to more impulsive or aggres-
sive clients (Beutler et al., 2004). Effective couple
therapists tend to be active enough to interrupt
destructive couple interaction, but not so active
that couples do not learn how to communicate
effectively on their own. Cultural and worldview
sensitivity has also been shown to be important
(Blow et al., 2007).
therapeutic alliance
The therapeutic alliance is the most studied
common factor in couple therapy. A strong
therapeutic alliance is generally seen as a necessary
component of effective therapy (Sprenkle et al.,
2009). Since the therapeutic alliance is more
complex in couple than in individual therapy, and
 common factors of change in couple therapy
it is so strongly related to outcome, we will discuss
the alliance in greater detail later.
Though the alliance is clearly related to outcome,
it is less clear whether that relationship is correla-
tional or causal. It is possible that if therapy is
proceeding well, clients may report a more positive
alliance. Clients may also report a more positive
alliance based on a third variable, such as changes
in his or her life outside of therapy, or the strength
of the referral source. Future research is needed to
determine the direction and strength of the rela-
tionship between alliance and outcome.
therapist/researcher allegiance
If a therapist or researcher strongly believes in the
approach he or she is using or studying, that belief
itself may increase the likelihood of success. This
passion likely spreads to the couple, increasing the
likelihood that the couple views the approach as a
credible path to relational health (Davis & Piercy,
2007a). People are more likely to “buy” something
the person “selling” it believes in. In other words,
the therapist may be more important than the
therapy (Blow et al., 2007).
Researchers conducting outcome studies typically
choose models they either developed or are passion-
ate about and typically recruit therapist participants
who share their passion, and that passion influences
outcome. In a study of psychotherapies, Luborsky
et al. (2002) coded allegiance effects three different
ways across 29 different outcome studies, finding
that “The correlation between the mean of 3
measures of the researcher's allegiance and the
outcome of the treatments compared was a huge
Pearson's r of .85 for a sample of 29 comparative
treatment studies” (p. 5). Wampold (2001) presents
similar evidence suggesting that what outcome
researchers think are treatment effects may rather
be allegiance effects.
organization and coherence provided
by a model
What is it about a therapeutic model that facil-
itates healing? We concur with Frank's assertion
(Frank & Frank, 1991) that the organization,
coherence, and healing rituals provided by the
model are far more important than the actual
content of the model. If a model provides ground-
ing in something the client's culture values as
relevant to healing (for couple therapy, scientifi-
cally based principles of healthy relationships),
provides credible healing rituals (i.e., interven-
tions), and a structure within which these rituals
make sense to the client, the model will likely help
clients be successful.
41
hope/expectancy
Although called by different terms, many empirically
validated couple or family therapy models focus
initial efforts largely on fostering hope (Sprenkle &
Blow, 2004). In couple therapy, the relationship of
hope to outcome is largely theoretical; we know of no
quantitative studies investigating the relationship
between hope and outcome. Ward and Wampler
(2010) and Davis and Piercy (2007b) used qualitative
methods to investigate hope. Ward and Wampler
define hope as “a belief and feeling that a desired
outcome is possible” (p. 216). Results of interviews
with several marriage and family therapists suggest
that hope consists of the partially overlapping
concepts of options (belief in one's ability to choose
a path to desired outcomes), evidence (can clients see
that the desired outcome is likely to occur?),
connection (hope instilled by a connection to other
human beings, a higher power, etc.), and actions
(one's willingness to act to reach desired outcomes).
Ward and Wampler discuss a general trajectory of
hope across successful treatment (from the therapist's
perspective), including pantheoretical interventions
couple therapists can use to facilitate hope at
different stages of treatment.
Davis and Piercy (2007b) found several concepts
related to hope in their interviews with couple
therapy model developers, their former students,
and their former clients. Trusted referral sources led
to more hope that clients could be helped, which led
them to give therapists more benefit of the doubt in
the crucial early sessions of therapy. This trust in
the referral source was amplified by the therapist's
actual competence, as defined by the ability of the
therapist to fit the model to each client's worldview
in a way that provided a credible road map out of
their problems. Motivational beliefs, metaphors,
and experiences were used throughout to help keep
the larger picture in mind when working through
difficult issues in therapy.
non-model-specific mechanisms
of change
Non-model-specific mechanisms of change exist
across different models regardless of model-specific
theoretical assumptions and techniques. They are
sometimes referred to as “narrow” (as a subset of
“broad” common factors, such as the alliance)
common factors (Sprenkle & Blow, 2004). These
mechanisms exist at a lower level of abstraction
than theory and a higher level of abstraction than
technique. For example, while narrative and
cognitive-behavioral couple therapists have differ-
ent conceptual lenses and employ different tech-
niques (e.g., re-storying and the downward arrow,
respectively), both achieve the same goal of helping
42 davis et al.
partners see each other in a “softer” light, which
will in turn help them treat each other more kindly.
“Changing the viewing” is the common change
mechanism. Sprenkle and Blow modified Karasu's
(1986) list of nonspecific change mechanisms to the
following three categories: changing the doing,
changing the thinking, and changing the feeling.
Common change mechanisms account for the
overlap among models in their natural form. If an
emotionally focused therapist, for example, is trying
to adapt his or her approach to a more cognitively
oriented couple, his or her work may start to look
very similar to a cognitive-behavioral therapist and
vice versa for a cognitive-behavioral therapist
working with a more affectively oriented couple.
We commend recent efforts by model developers to
discuss their models in more pluralistic terms, and
believe that the more this happens, the easier it will
be for therapists to master more than one approach
and ultimately better adapt to a diverse clientele.
Dattilio (2010), for example, goes to great lengths
to incorporate emotions into a cognitive-behavioral
framework, helping his approach resonate with a
wider range of therapists.
Common Factors Unique to Couple Therapy
Many of the common factors reviewed above apply
to individual therapy as well as couple therapy.
There may be a few important differences, however,
between common factors in individual and couple
therapy (Sprenkle et al., 2009).
conceptualizing difficulties in
relational terms
Many symptoms become understandable responses
to untenable circumstances when a therapist steps
back and looks at a couple's larger context. If a
therapist only heard the husband's story about his
nagging wife, the therapist may ultimately side with
the husband, downplaying his role in the couple's
distress and unfairly demonizing the wife. Such an
approach could unwittingly undermine the mar-
riage and subtly lead the husband toward separa-
tion or divorce. However, once the therapist
brought the wife into treatment, her nagging will
likely become an understandable response to the
husband's withdrawing, for example. These recip-
rocal interactions in which one person's symptoms
become understandable only in the context of their
partner's symptoms, and each are presumed to
exacerbate and maintain the other, is a hallmark of
systemic couple therapy (with possible exceptions,
such as domestic violence).
While there is little direct evidence for this common
factor (i.e., comparing relationship conceptualiza-
tion to individual conceptualization, which is not the
same as comparing conjoint vs. individual treat-
ments), the clear effectiveness of systemic approaches
that embrace this common factor relative to individ-
ual approaches provides strong indirect evidence.
Some issues reviewed earlier achieve a more durable
resolution in systemic treatment. In their review of 20
meta-analyses of couple and family therapy, Shadish
and Baldwin (2002) note that the average effect size
for couple therapy is 0.84, with the average for
family therapy being 0.58. The couple therapy effect
size is strong (stronger than many medical treat-
ments); the family therapy effect size is moderate,
possibly due to the difficulty of getting everyone in a
family happy about a particular outcome or the
difficulty of problems typically addressed in family
therapy.
disrupting dysfunctional relational
patterns
If conceptualizing problems in terms of relational
patterns is a common factor of effective couple
therapy, it follows that disrupting those patterns is
also a common factor. Similar indirect evidence
exists for this common factor as exists for concep-
tualizing problems in relational terms. Davis and
Piercy (2007a, 2007b) found that therapists using
different couple therapies described conceptualizing
problems in terms of interactional cycles (i.e.,
relational patterns), and focused the majority of
their interventions on altering these patterns.
Furthermore, interactional cycles were found to
consist of cognitive, behavioral, and affective
domains, all of which were targets of intervention.
For example, a wife thinks her husband does not
care about her when he comes home late (cognition)
so she gets angry (affect) and yells at him every
time he is late (behavior). Her husband thinks his
wife is a nag (cognition) and feels intimidated by her
(affect) so he comes home late each night (behav-
ior). A therapist would have six possible points
of intervention into this interactional cycle, all of
which could shift the entire pattern. If the wife
realized that she felt abandoned rather than angry,
and she was able to express that to her husband
rather than nagging, he may soften toward her and
stop thinking she's a nag and start coming home on
time, to which she may respond more kindly. Or if
the therapist helped the husband start coming home
on time no matter what, it may soften the wife's
thoughts about and feelings toward him, which
would shift the cycle, and so forth. While some
models focus more on one aspect of the cycle than
others (e.g., EFT focuses on emotions more than
CBT, which focuses more on cognitions and
behaviors), most systemic models focus on each
aspect to a degree.
 common factors of change in couple therapy
expanded direct treatment system
A systemically oriented therapist typically works with
as many members of a client's system as possible. If
working with an individual, the same therapist will
still incorporate a thorough understanding of the
client's systems into assessment, conceptualization,
and intervention. Symptoms are seen as best under-
stood in the context of the client's interactions with
micro (e.g., spouse or family) and/or macro (e.g.,
larger cultural and political) systems. The more
people in the system who change, the more durable
that change is likely to be. Greater efficacy for several
problems discussed earlier provides substantial direct
evidence for this common factor.
expanded therapeutic alliance
There is ample evidence supporting the therapeutic
alliance as a potent common factor. A couple
therapist must monitor not only his or her alliance
with each partner, but their alliance with each other
and the overall alliance as a group working
together. This is particularly challenging when
working in the highly emotional climate that so
often characterizes couple therapy in which one or
both partners is prone to interpret therapist
attempts to validate the partner as a personal
rejection. When the alliance is strong, it tends to
amplify the effects of treatment. When it is weak,
therapy often does not last long enough for
interventions to take effect (Sprenkle et al., 2009).
Therefore, monitoring, maintaining, and repairing
the alliance is a critical task of couple therapy.
The alliance is more than the emotional “bond”
between clients and therapist, including also tasks
and goals (Horvath, 1994). In addition to feeling
emotionally connected to their therapist (bond),
clients and therapist have to agree regarding where
they are headed (goals) and how to get there (tasks).
In relational therapy, the relationship is the client
rather than the individuals within the relationship.
This requires nuanced skills at balancing the
alliance so each partner feels validated as the
therapist listens to them, yet not threatened while
the therapist validates his or her partner, who is
often saying inflammatory things. There is some
research suggesting that balancing the alliance is
even more important than the strength of the
alliance (Sprenkle & Blow, 2004). The alliance in
couple therapy is very complex. Space limitations
prevent a thorough discussion of all the nuances of
forming, monitoring, maintaining, and repairing
the alliance in couple therapy. Brief practical
guidance for maintaining a systemic alliance is
found later in the “Clinical Implications” section.
Interested readers are referred to Sprenkle et al.
(2009) for a more thorough review.
43
Extreme Versus Moderate Common Factors
Approaches
Although there is general agreement on what
common factors are, there is disagreement about
the clinical and research implications of common
factors. Many scholars have taken what we consider
an extreme “either/or” stance on common factors.
For example, it is common to hear “All models work
the same; therefore models do not matter,” “A strong
therapeutic alliance is all that is needed for success,”
or “Because meta-analyses reveal that no significant
differences between treatments exist, comparative
efficacy research is useless and should stop.” We
reject these and similar extreme statements as reach-
ing beyond what the data justify, and believe they run
the risk of throwing the proverbial baby out with the
bath water. Our “both/and” moderate common
factors approach (Sprenkle et al., 2009) takes a
more centrist stance, and draws five important
distinctions.
belief that one treatment is as good
as another versus claims of relative
efficacy
Many common factors proponents cite meta-
analytic research in which no model has consis-
tently shown to be more efficacious than another
(Shadish & Baldwin, 2002) as evidence that all
models are the same. “The dodo bird verdict”—a
phrase borrowed from Alice in Wonderland in
which all have won and must have prizes—is often
used to make this point (Luborsky et al., 2002).
Taken to the extreme, the dodo bird renders models
irrelevant, and implies that it does not matter what
you do in therapy provided you have a good
alliance, a motivated client, and so forth.
We believe the data support claims that among
efficacious couple therapies, there is little evidence for
differences in treatment outcome (Shadish & Baldwin,
2002). However, just because tested models have not
been shown to be differentially or relatively efficacious
vis-à-vis each other does not mean they are irrelevant.
On the contrary, most tested couple therapy models
show significant improvement, and differences in
outcome between the model group and treatment as
usual group in most RCTs suggest that the model adds
something essential to treatment (Shadish & Baldwin,
2002). Using a well-defined, tested couple therapy
approach such as EFT (Johnson, 2004) ensures a
consistently applied effective therapy, whereas the
impact of untested or atheoretical “improvisational”
therapy remains unknown. It may be that some of
those therapies do not include the core effective
ingredients of successful couple therapy.
We believe it likely from the findings of meta-
analyses (Shadish & Baldwin, 2002) that successful
44 davis et al.
results will also emerge from many (but not all) as
yet untested couple therapies such as narrative
therapy and others. To the extent that such theories
are based on sound psychological and systemic
principles, provide adequate organization and
coherence, share common elements, and are imple-
mented by competent therapists, they would likely
(but not certainly) be proven efficacious. Rarely is a
model meeting those criteria for bona fide treat-
ments found not to be efficacious in clinical trials
(Wampold, 2001). Nevertheless, we remain open to
the possibility that differences may be discovered as
research methodologies and questions become
more refined.
disparages rather than supports
treatment models
We believe that well-defined models are an indis-
pensible component of good therapy. Models help a
therapist know how and when to intervene and to
what purpose. Models provide necessary order and
structure to therapy. Models are like a road map that
directs therapists through the maze of information
present in therapy and orients therapists and clients
to what is wrong, proposes viable alternatives for
helping things go right, and techniques, skills, and
rituals for helping clients go from one to the other as
they adapt the therapy's offerings to their needs in
their own way. Our only contention is with the claim
that models work primarily through mechanisms
that are unique to that model. Rather, we believe
effective models are the vehicles through which
common factors operate.
primarily emphasizes the therapeutic
relationship rather than views the
relationship as only one aspect
of change
Common factors proponents often say “The
therapeutic alliance is all that matters.” We believe
that while the therapeutic alliance is a necessary
factor of successful therapy, it is not sufficient, in
part because it does not stand separate from other
variables in therapy. For example, whether the
client views a therapist's theory and interventions as
credible has an influence on the alliance (Sprenkle
& Blow, 2004).
minimizes rather than supports
clinical trials research
As discussed earlier, we are more optimistic about the
usefulness of clinical trials research than some of our
common factors colleagues whose work we other-
wise admire (Wampold, 2001). Not only do clinical
trials establish the efficacy of therapy to third-party
stakeholders—an important contribution, even if
largely political—they also form the foundation for
the meta-analytic studies that led to the common
factors paradigm. Furthermore, while many com-
mon factors cannot be experimentally manipulated,
some (e.g., therapist factors, client motivation,
therapist belief in and degree of adherence to the
model) could easily be added to the list of variables
measured in a clinical trial. While many of these
variables are best suited for quantitative and
qualitative process research designs, clinical trials
could still provide useful information about common
factors. Again, we stress the need not to throw the
baby out with the bath water.
“either–or” versus “both–and”
As discussed earlier, we are open to contemporary
common factors claims being subject to future
refinement or disconfirmation as research method-
ologies become more complex. Some of this is
already happening; in individual therapy there
appears to be reasonable evidence that certain
specific problems such as phobias, panic disorder,
and compulsions respond better to behavioral and
cognitive approaches, and that these results cannot
be explained by the usual mediating and moderat-
ing variables in outcome research (Sprenkle &
Blow, 2004). Similarly, certain types of family
problems respond better to certain types of therapy,
though there are many similarities among the
effective therapies for this population. Similarly,
couple therapies that have been shown to be
effective are all structured, include a caring,
involved therapist who monitors and works with
the interactions, builds skills, promotes emotional
engagement, provides a calm holding environment
for dealing with negative highly charged feelings,
promotes communication, and targets the building
of the relationship system. We believe that such
common factors account for more variance than
specific factors, though we acknowledge this may
not be the case for every population.
A Common Factors Couple Therapy
Metamodel
The common factors movement is sometimes
criticized as providing little more than lists of
variables without any guidance as to how and when
these variables interact to produce change (Sexton
& Ridley, 2004). Similar to the principle- and
process-based approaches discussed earlier, a com-
mon factors metamodel could address this problem
by outlining a framework of conceptual principles
and mechanisms of change that could be super-
imposed over diverse clinical approaches, allowing
therapists to move between models as indicated by
general principles and client needs.
 common factors of change in couple therapy
Davis and Piercy (2007a, 2007b) used qualitative
grounded theory methodology to develop a prelim-
inary metamodel. Diverse couple therapy model
developers and proponents, their former students,
and former clients of both were asked what they
thought made therapy successful. Participants
described couples as entering therapy without a
clear problem definition or a workable plan for
resolving their problems, leading them to feel
hopeless. Conversely, the therapist described hav-
ing hope and direction in the form of their therapy
model and experience. Clients reported increased
hope as they adopted their therapist's way of
viewing their situation. This adoption seemed to
happen provided that (a) the therapist's explana-
tion of their difficulties fit the client's experience,
(b) the therapist was viewed as competent (which
seemed to be related to the first concept), (c) the
therapist proposed ideas in a way that minimized
client resistance, and (d) the clients were at least
somewhat willing to take responsibility for their
problems.
Though most interventions were unique to the
therapist's model, they seemed to follow similar
patterns (Davis & Piercy, 2007a, 2007b). For
example, therapists described helping clients
become aware of their interactional cycles by
slowing down the cycle, thus helping each partner
step away from their struggles to change each
other. As couples slowed down the cycle,
therapists reported helping clients stand “above”
the cycle, thus seeing their and their partner's role
in the cycle in a different light. Experiencing their
partner differently seemed to make it easier for
each partner to take responsibility for their role in
their difficulties and take the recommended steps
to exit the cycle. As therapy neared completion,
clients described a softening toward each other.
They also said they stopped trying to control,
choosing instead to enjoy each other regardless of
differences. They described feeling more confident
and relaxed. These processes are similar to
softening in EFT (Johnson, 2004) and acceptance
in IBCT (Jacobson & Christensen, 1996). Partic-
ipants described changes at different stages of
therapy as being mediated by therapist and client
variables, the alliance, expectancy/motivational
factors, and other variables as operationalized in
Davis and Piercy.
There are several limitations inherent in the
design of the Davis and Piercy (2007a, 2007b)
study. First, the goal was an in-depth description of
a process rather than a statistical investigation of
causal links between variables. Furthermore, the
sample was a small, carefully chosen group of
highly skilled practitioners who recruited successful
45
clients. Participants also focused on their subjective
memory of therapy rather than objective observa-
tions of the therapy by trained coders. Despite the
limitations inherent in this study specifically, and in
qualitative research generally, we believe that
formal inductive inquiry is the best way to generate
relevant hypotheses during the formative stages of
developing a new approach such as a common
factor metamodel.
Training, Clinical, and Research Implications of
a Moderate Common Factors Approach
clinical implications
Providing clinical guidance for each common factor
is beyond the scope of this article. Interested readers
can find such guidance in Sprenkle et al. (2009). We
limit our discussion to the therapeutic alliance
because it is one of the most robust yet complex
common factors in couple therapy. Boszormenyi-
Nagy coined the term “multidirectional partiality”
(Boszormenyi-Nagy & Spark, 1984) to describe a
process whereby the therapist actively takes each
partner's side, pleading his or her case to his or her
partner in language that is easier to hear. By taking
each person's side, both partners feel accepted and
start to entertain the possibility that there is validity
to what their partner is saying. As the therapist
reframes each partner's comments in ways that are
easier to hear, they learn to say things without
attacking and listen without defending. It is
important to spend roughly equal session time
with each partner during this process. When it is
necessary to focus more on one partner, it can be
helpful to clarify that the relationship is the client,
so for the good of the relationship the focus may be
momentarily imbalanced. While the balance of
attention can vary from session to session, it is still
important to keep the alliance balanced over the
course of therapy.
Split alliances, in which the therapeutic alliance is
stronger with one partner than the other, present
another challenging situation. Shifting from the
engaged to disengaged partner too abruptly may
end therapy prematurely. However, therapy may
not progress unless the disengaged partner engages.
Shifting positions solely for the sake of the alliance
is likely to fail except in the case of ethical or social
justice issues. Rather, therapists are encouraged to
use dialogue, empathy, and styles of relating that
engage the disengaged partner while still validating
the engaged partner.
training implications
Given the importance of adapting an approach to a
client's culture and worldview, it follows that a
therapist should achieve competence in diverse
46 davis et al.
models. We reiterate what Davis and Sprenkle have
said elsewhere with Blow:
“[Common factors ideas] suggest a shift in our
thinking away from encouraging a student to be
passionate about a theory towards being passio-
nate about theory. We believe that encouraging a
therapist to be passionate about only one theory
may unwittingly give a therapist the proverbial
hammer with which he or she must turn every
client into a nail before treatment can proceed. We
advocate that a more client-centered approach
would be to encourage the therapist to have a
thorough familiarity with several diverse models
so that he or she can adapt to his or her client's
contexts rather than vice-versa. . . . Clients should
not have to add ‘figure out how to adapt to my
therapist’ to their already lengthy list of chal-
lenges.” (Blow et al., 2007, pp. 309–310, emphasis
in original)
As reviewed earlier, whereas more research is
needed to substantiate this claim, indirect evidence
suggests that therapists who tailor their treatment
based on client feedback enjoy significantly better
results (Halford et al., 2012-this issue). The ability
to do this presupposes the knowledge of different
models, or at least the ability to present the same
model in widely varying ways.
The daunting task of learning seemingly diverse
models can be made easier with an understanding
of the nonspecific mechanisms of change discussed
earlier. When a therapist learns to view models as
more overlapping than distinct, moving between
approaches can become more intuitive (Blow et al.,
2007). Realizing the overlap among seemingly
disparate models can also help therapists become
passionate about diverse models, thus increasing
their chances of success. In fact, it could be argued
that many models of couple therapy were devel-
oped by people who had a thorough grasp of and
appreciation for diverse models. Examples include
EFT, which draws upon structural–strategic and
experiential therapies (Johnson, 2004) and IBCT,
which incorporates emotionally focused principles
into a cognitive-behavioral framework (Jacobson
& Christensen, 1996).
We also believe that therapists should receive
specific training in common factors such as establish-
ing, maintaining, and repairing the therapeutic
alliance, seeking and responding appropriately to
client feedback, nurturing hope, and facilitating
motivation. Furthermore, we believe training pro-
grams should more deliberately help students develop
personal therapeutic attributes. State licensure boards
could also require similar training in the continuing
education requirements for license renewal.
research
We suggest there is a need for far greater focus
in research on the common factors in effective
treatment. Clinical trials research can be quite
useful in this regard, and could be greatly enhanced
by incorporating a wider variety of independent
variables discussed earlier (e.g., therapist and client
variables) into the design. The ability of process/
observational and outcome research to shed light
on the how of change in couple therapy make them
promising research designs for the common factors
paradigm. In addition to the research directions
suggested at the beginning of this article, qualitative
methodologies also could be useful, particularly
because there are so many unanswered questions
in the current common factors paradigm that at
this point could benefit from a discovery form of
investigation. Such designs could shed light on
largely unanswered yet critical questions such as
“How do therapists balance the alliance in a way
that allows disparate views to be validated simul-
taneously?” “What client characteristics are related
to outcome and how can a therapist influence
them?” “What therapist behaviors or characteris-
tics facilitate versus thwart client motivation?”
“What—if any—common change mechanisms are
indicated for different problems (e.g., in what ways,
if any, is helping a couple through an affair different
than helping them through domestic violence?)?”
and so forth.
Conclusion
We believe that the moderate common factors
paradigm provides the best explanation of the data
suggesting few differences between couple therapy
approaches (Shadish & Baldwin, 2002). While
much of the relevant research into common factors
in couple therapy remains to be done, we never-
theless believe there is enough to suggest a strong
basis for the ideas we have covered in this article.
We hope that research into common factors in
couple therapy will continue to gain momentum.
References
Anker, M. G., Duncan, B. L., & Sparks, J. A. (2009). Using
client feedback to improve couple therapy outcomes: A
randomized clinical trial in a naturalistic setting. Journal of
Consulting and Clinical Psychology, 77, 693–704.
Atkins, D. C., Berns, S. B., George, W., Doss, B., Gattis, K., &
Christensen, A. (2005). Prediction of response to treatment
in a randomized clinical trial of marital therapy. Journal of
Consulting and Clinical Psychology, 73, 893–903.
Beutler, L. E., Malik, M. L., & Alimohamed, S. (2004).
Therapist variables. In M. J. Lambert (Ed.). Bergin and
Garfield's handbook of psychotherapy and behavior change
(pp. 227–306). New York: Wiley.
 common factors of change in couple therapy
Blatt, S. J., Sanislow, C. A., Zuroff, D. C., & Pilkonis, P. A.
(1996). Characteristics of effective therapists: Further
analyses of data from the National Institute of Mental
Health treatment of depression collaborative research
program. Journal of Consulting and Clinical Psychology,
64, 1276–1284.
Blow, A. J., Sprenkle, D. S., & Davis, S. D. (2007). Is who
delivers the treatment more important than the treatment
itself?: The role of the therapist in common factors. Journal
of Marital and Family Therapy, 33, 298–317.
Bodenman, G., Plancherel, B., Widmer, K., Meuwly, N.,
Hautzinger, M., Beach, S. R. H., Gabriel, B., et al. (2008).
Effects of coping-oriented couples therapy on depression: A
randomized clinical trial. Journal of Consulting and Clinical
Psychology, 76(6), 944–954.
Bohart, A. C., & Tallman, K. (2010). Clients: The neglected
common factor in psychotherapy. In B. L. Duncan, S. D.
Miller, B. E. Wampold, & M. A. Hubble (Eds.). The heart
and soul of change: Delivering what works in therapy
(2nd ed., pp. 83–112). Washington, DC: American
Psychology Association.
Boszormenyi-Nagy, I., & Spark, G. M. (1984). Invisible
loyalties. New York: Routledge.
Chambless, D. L., & Ollendick, T. H. (2001). Empirically
supported psychological interventions: Controversies
and evidence. Annual Review of Psychology, 52,
685–716.
Christensen, A. J. (2010). A unified protocol for couple therapy.
In K. Halweg, M. Grawe-Gerber, & D. H. Baucom (Eds.).
Enhancing couples. Cambridge, MA: Hogrefe.
Cline, V. B., Mejia, J., Coles, J., Klein, N., & Cline, R.
(1984). The relationship between therapist behaviors
and outcome for middle- and lower-class couples in
marital therapy. Journal of Clinical Psychology, 40,
691–704.
Dattilio, F. M. (2010). Cognitive-behavioral therapy with
couples and families: A comprehensive guide for clinicians.
New York: Guilford Press.
Davis, S. D., & Piercy, F. P. (2007). What clients of MFT
model developers and their former students say about
change, Part I: Model dependent common factors across
three models. Journal of Marital and Family Therapy, 33,
318–343.
Davis, S. D., & Piercy, F. P. (2007). What clients of MFT model
developers and their former students say about change, Part
II: Model independent common factors and an integrative
framework. Journal of Marital and Family Therapy, 33,
344–363.
Doss, B., Thum, Y., Sevier, M., Atkins, D., & Christensen, A.
(2005). Improving relationships: Mechanism of change in
couple theraphy. Journal of Consulting and Clinical
Psychology, 73, 624–633.
Elkin, I., Shea, T., Watkins, J. T., Imber, S. D., Sotsky, S. M.,
Collins, J. F., et al. (1989). National Institute of Mental
Health treatment of depression collaborative research
program: General effectiveness of treatments. Archives of
General Psychiatry, 46, 971–982.
Frank, J. D., & Frank, J. B. (1991). Persuasion and healing: A
comparative study of psychotherapy (3rd ed.). Baltimore:
Johns Hopkins University Press.
Halford, W. K., Hayes, S., Christensen, A., Lambert, M.,
Baucom, D. H., & Atkins, D. (2012). Towards making
progress feedback an effective common factor in couple
therapy. Behavior Therapy, 43, 49–60 (this issue).
Halford, W., & Snyder, D. (2012). Special series: Universal processes
and common factors in couple therapy and relationship
education-Introduction. Behavior Therapy, 43, 1–12 (this issue).
47
Horvath, A. O. (Ed.). (1994). Empirical validation of Bordin's
pantheoretical model of the alliance: The Working Alliance
Inventory perspective. Oxford, England: Wiley.
Jacobson, N. S., & Christensen, A. (1996). Integrative couple
therapy: Promoting acceptance and change. New York:
Norton.
Johnson, S., & Greenberg, L. (1985). The differential effective-
ness of experiential and problem solving interventions in
resolving marital conflict. Journal of Consulting and
Clinical Psychology, 53, 175–184.
Johnson, S. M. (2004). The practice of emotionally focused
marital therapy: Creating connection (2nd ed.). New York:
Brunner/Mazel.
Johnson, S. M., & Talitman, E. (1997). Predictors of success in
emotionally focused marital therapy. Journal of Marital and
Family Therapy, 23, 135–152.
Karasu, T. B. (1986). The specificity versus nonspecificity
dilemma: Toward identifying therapeutic change agents.
American Journal of Psychiatry, 143, 687–695.
Lambert, M. J., Whipple, J. L., Smart, D. W., Vermeersch, D. A.,
Nielsen, S. L., & Hawkins, E. J. (2001). The effects of
providing therapists with feedback on patient progress during
psychotherapy: Are outcomes enhanced? Psychotherapy
Research, 11, 49–68.
Luborsky, L., Rosenthal, R., Diguer, L., Andrusyna, T. P.,
Berman, J. S., Levitt, J. T., et al. (2002). The dodo bird
verdict is alive and well—mostly. Clinical Psychology:
Science and Practice, 9, 2–12.
McCrady, B. S., Epstein, E. E., Cook, S., Jensen, N., &
Hildebrandt, T. (2009). A randomized trial of individual
and couple behavioral alcohol treatment for women. Journal
of Consulting and Clinical Psychology, 77(2), 243–256.
Orlinsky, D. E., Grawe, K., & Parks, B. K. (1994). Process and
outcome in psychotherapy: Noch einmal. In A. E. Bergin, &
S. L. Garfield (Eds.). Handbook of psychotherapy and
behavior change (4th ed., pp. 270–376). Oxford, England:
Wiley.
Powers, M. B., Vedel, E., & Emmelkamp, P. M. G. (2008).
Behavioral couples therapy (BCT) for alcohol and drug use
disorders: A meta-analysis. Clinical Psychology Review,
28(6), 952–962.
Prochaska, J. O., & Norcross, J. C. (2009). Systems of
psychotherapy: A transtheoretical analysis (7th ed.). New
York: Brooks Cole.
Sexton, T. L., & Ridley, C. R. (2004). Implications of a
moderated common factors approach: Does it move the
field forward? Journal of Marital and Family Therapy, 30,
159–164.
Shadish, W. R., & Baldwin, S. A. (2002). Meta-analysis of
MFT interventions. In D. H. Sprenkle (Ed.). Effectiveness
research in marriage and family therapy (pp. 339–370).
Alexandria, VA: American Association of Marriage and
Family Therapy.
Smith, M. L., & Glass, G. V. (1977). Meta-analysis of
psychotherapy outcome studies. American Psychologist,
32(9), 752–760.
Snyder, D. K. (1999). Affective reconstruction in the context of
a pluralistic approach to couples therapy. Clinical Psychology:
Science and Practice, 6, 348–365.
Sprenkle, D. H., & Blow, A. J. (2004). Common factors and our
sacred models. Journal of Marital and Family Therapy, 30,
113–130.
Sprenkle, D. H., Davis, S. D., & Lebow, J. (2009). Common
factors in couple and family therapy: The overlooked
foundation for effective practice. New York: Guilford Press.
Stith, S. M., Rosen, K. H., McCollum, E. E., & Thomsen, C. J.
(2004). Treating intimate partner violence within intact
48 davis et al.

couple relationships: Outcomes of multi-couple versus
individual couple therapy. Journal of Marital and Family
Therapy, 30(3), 305–318.
Stolk, Y., & Perlesz, A. J. (1990). Do better trainees make worse
family therapists?: A follow-up study of client families.
Family Process, 29, 45–58.
Wampold, B. E. (2001). The great psychotherapy debate:
Models, methods, and findings. Mahwah, NJ: Erlbaum.
Ward, D. B., & Wampler, K. S. (2010). Moving up the
continuum of hope: Developing a theory of hope and
understanding its influence in couples therapy. Journal of
Marital and Family Therapy, 36, 212–228.
R E C E I V E D : August 17, 2010
A C C E P T E D : January 5, 2011
Available online 6 June 2011