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

net/publication/298723116

Common (Nonspecific) Factors in Psychotherapy

Chapter · January 2015

DOI: 10.1002/9781118625392.wbecp272

CITATIONS READS
34 13,770

2 authors:

Marcus Huibers Pim Cuijpers

221 PUBLICATIONS 11,017 CITATIONS

Vrije Universiteit Amsterdam
1,447 PUBLICATIONS 108,226 CITATIONS
SEE PROFILE
SEE PROFILE

All content following this page was uploaded by Marcus Huibers on 05 December 2017.

The user has requested enhancement of the downloaded file.

Common (Nonspecific) have prizes.” This emphasizes that despite the
apparent differences between psychotherapies,
Factors in Psychotherapy they are all effective and lead to comparable
Marcus J. H. Huibers and Pim Cuijpers outcomes. In the decades that followed, the
VU University Amsterdam, The Netherlands finding of equivalent outcomes was repeatedly
demonstrated in reviews and meta-analyses,
although not all meta-analyses support this
Although numerous trials and meta-analyses
conclusion (Tolin, 2010). To this date, the dodo
have demonstrated the effectiveness of psy- bird verdict has been an issue for controversy
chotherapy for a variety of psychological and debate, as described below.
problems, the mechanisms by which positive In his influential book Persuasion and Heal-
psychotherapeutic change occurs are less cer- ing: A Comparative Study of Psychotherapy
tain. There are many forms of psychotherapy, (1961), Jerome Frank (1889–1957) described
and most of them offer a particular theory four essential elements common to various
to explain the process of psychotherapeutic forms of mental healing, including psy-
change. Theoretical differences among psy- chotherapy: a functioning relationship between
chotherapies suggest differential pathways patient and therapist; a rationale that provides
of change. That is, particular elements or credibility to the treatment being delivered;
techniques of a given type of psychotherapy certain procedures or rituals that are provided
are presumed to account for any observed in a structured manner; and a healing context
therapeutic change. or setting.
However, despite the enormous number of Lambert and Ogles (2004) describe the
distinct psychotherapies, many therapies pro- potential explanations for the purported
duce equivalent outcomes, as meta-analyses finding that all psychotherapies produce
have shown persistently for many decades comparable outcomes: (a) different forms of
(Lambert & Ogles, 2004). This phenomenon, psychotherapy can lead to comparable out-
known as the dodo bird verdict, has led some comes through different processes; (b) some
authors to suggest that factors common to forms of psychotherapy are in fact supe-
all forms of psychotherapy (e.g., therapeutic rior to other therapies, but these differences
relationship) are largely responsible for posi- have not been adequately detected in past
tive psychotherapeutic change. A great deal of treatment studies; or (c) different forms of
research focuses on the role of common factors, psychotherapy embody common factors that
in the psychotherapeutic change process. are curative, though not emphasized by the
The so-called common factor model was first theory of change of that particular form of
introduced by Saul Rozenzweig (1907–2004) psychotherapy. They describe the different
in 1934. Making more an observation than an conceptualizations of the term common fac-
actual review, Rozenzweig noted that all psy- tors, and provide an overview of the common
chotherapies produced equivalent outcomes, factors that can be derived from the literature.
which means that they must work through As can be seen in Table 1, they are grouped into
factors common to all of them. Paraphrasing support, learning, and action categories, based
from Lewis Carroll’s novel Alice’s Adventures on the sequential order of change that can
in Wonderland, he coined the term dodo bird be observed in psychotherapy. For example,
verdict, following the dodo bird’s pronounce- a strong therapeutic alliance can lead to a
ment that “Everybody has won and all must corrective emotional experience, which then

The Encyclopedia of Clinical Psychology, First Edition. Edited by Robin L. Cautin and Scott O. Lilienfeld.
© 2015 John Wiley & Sons, Inc. Published 2015 by John Wiley & Sons, Inc.
DOI: 10.1002/9781118625392.wbecp272
2 COMMON (NONSPECIFIC) FACTORS IN PSYCHOTHERAPY

Table 1 Sequential listing of common factors.

Support factors Learning factors Action factors

Catharsis Advice Behavioral regulation

Identification with therapist Affective experiencing Cognitive mastery
Mitigation of isolation Assimilating problematic Encouragement of facing fears
experiences
Positive relationship Cognitive learning Taking risks
Reassurance Corrective emotional experience Mastery efforts
Release of tension Feedback Modeling
Structure Insight Practice
Therapeutic alliance Rationale Reality testing
Therapist/client active participation Exploration of internal frame of Success experience
reference
Therapist expertness Changing expectations of Working through
personal effectiveness
Therapist warmth, respect, empathy,
acceptance, genuineness
Trust

leads the patient to facing his or her fears or Several meta-analyses show that a strong
dealing with his or her problems. As Lambert therapeutic alliance is linked to treatment
and Ogles acknowledge, this sequential process success in psychotherapy, although the corre-
of change is a theoretical assumption that has lation is modest, with only 7% of the variance
yet to be tested in empirical research. explained (Fluckiger, Del Re, Wampold,
Much cited in the professional literature Symonds, & Horvath, 2012). Findings like
for active clinicians is a pie chart that was these have consistently been presented as evi-
developed by Lambert (1992) in which he dence for the common factor model, although
summarizes the therapy factors that determine correlation data cannot be used to infer
the outcome in psychotherapy. He claims that causation.
40% of the change can be attributed to factors Emphasis on the alliance has developed
outside the therapy, 15% to the placebo effect originally out of psychodynamic psychother-
(expectation of improvement), 15% to factors apy, but is nowadays accepted by all forms
that are specific to the type of psychother- of psychotherapy; although the hypothesized
apy delivered, and 30% to common factors, role of alliance may differ among treatment
emphasizing the importance of common fac- approaches. As with common factors in gen-
tors over specific factors and placebo effects. eral, therapeutic alliance of relationship has
It should be noted however, that this pie chart been defined in many different ways that
is not based on actual empirical evidence on will not be described here. Gaston (1990)
causal mechanisms of change, but on an aggre- reviewed the predominantly psychodynamic
gated estimation of observational correlations and client-centered literature on the concept
found in the literature. It is therefore merely of alliance, and proposed the following four
hypothetical, despite its appeal to professional components: (a) the therapeutic alliance,
workers. or the patient’s affective relationship to the
Of all the identified common factors, the therapist; (b) the working alliance, or the
therapeutic alliance or relationship between patient’s ability to purposefully work in ther-
therapist and patient is undisputedly the most apy; (c) the therapist’s empathic understanding
prominent and most extensively researched. and involvement; and (d) patient–therapist
 COMMON (NONSPECIFIC) FACTORS IN PSYCHOTHERAPY 3

agreement on the goals and tasks of treat- treatments in specific patient groups remain
ment. This conceptualization has been widely undetected due to limitations in the research
used ever since. The uniform measurement of design (e.g., insufficient statistical power).
alliance with instruments such as the Work- DeRubeis et al. also point to the danger of
ing Alliance Inventory (WAI) that tab into interpreting the (modest) correlation between
these components has greatly advanced the therapeutic alliance and outcome as causal
comparative research in this field. evidence in favor of the common factor the-
ory. To be able to draw such a conclusion, a
The Great Debate on Mechanisms temporal relation must be demonstrated in
of Change which change in the alliance actually precedes
change in symptoms. Most alliance studies do
Ever since Rozenzweig coined the term dodo not have a temporal design that allows for such
bird verdict, a heated debate has persisted on conclusions. It might therefore very well be that
the mechanisms that account for the effects symptom change leads to a better therapeutic
of psychotherapy. Although almost everyone alliance (reversed causation), not the other
agrees that both common and specific factors way around. Repeated measurements of pro-
are active in psychotherapy, advocates of the cesses and outcome throughout therapy and
specific factors model emphasize the specific mediation analyses that rule out a temporal
ingredients of the different psychotherapies confound are needed to evaluate the impor-
thought to be responsible for their observed tance of the alliance as a curative factor. And
effectiveness, whereas opponents claim that even if a temporal relation were found, this
common factors, such as the therapeutic association would still only be correlational,
relationship, are pivotal. In a special issue and would not rule out the possibility that
of Clinical Psychology: Science and Practice, a third, unmeasured, variable is responsible
prominent voices debated this issue. for the change in the alliance as well as in the
DeRubeis, Brotman, and Gibbons (2005) outcome.
criticize the common factor model by ques- Craighead, Sheets, and Bjornsson (2005),
tioning the lines of research that are often although largely agreeing with DeRubeis et al.
cited in support of the common factor model. (2005), criticize the stark dichotomy made
They argue that the claim of outcome equiv- between common and specific factors. Thera-
alence among psychotherapies neglects those peutic outcomes, they argue, are not the sole
comparative outcome studies that do show function of only specific or common factors.
that a particular psychotherapy is superior Superior effects of one treatment over another
to another form, such as the superiority of do not prove that specific effects are respon-
cognitive behavioral therapy (CBT) for panic sible for the observed superiority. Most likely,
disorder. Moreover, even if psychotherapies do therapists in studies that have demonstrated
produce comparable outcomes, this does not superiority maximized common factor effects
prove that common factors are responsible. It such a strong therapeutic alliance, a credible
is possible that different psychotherapies work treatment rationale, and so on. Craighead et al.
through specific and different ways to produce also note that the “sudden gains” (i.e., dramatic
change. Antidepressants and psychotherapy symptom improvements between sessions
are known to produce broadly equivalent early on in treatment) that DeRubeis et al.
effects, but in this case we accept the notion attribute to specific CBT interventions can also
that they most likely (although not necessarily) be observed in pharmacological treatments,
work through differential (i.e., biological ver- suggesting that these sudden improvements
sus psychological) pathways. It is also possible can also result from common factors such
that the number of possible mediators and as remoralization and hope. Specific factors,
moderators is so large that differences among such as therapeutic techniques, and common
4 COMMON (NONSPECIFIC) FACTORS IN PSYCHOTHERAPY

factors, such as alliance, probably go together studies in which the proposed common (or
and interact, thus adding to the full effects specific) factor is isolated and directly manip-
of psychotherapy. This opinion is shared by ulated, so that a causal link to outcome can
Castonquay and Grosse Holtforth (2005), who be established. Unfortunately, such studies are
state that, given the many potential factors that scarce, if not nonexistent. In order to clarify
are involved in the process, one is forced to the mechanisms responsible for therapeutic
recognize that the complexity of psychother- change, experimental manipulation of isolated
apy goes beyond a debate between common factors is needed.
versus specific factors, or techniques versus
relationship. The Role of the Therapeutic
Wampold (2005), on the other hand, returns Alliance: Differential Effects
to the issue of outcome equivalence, and for Different Treatments?
essentially claims the common factor model
to be an evidence-based model. In response Most studies on common factors such as the
to DeRubeis et al. (2005), Wampold states alliance have been merely correlational in
that the null hypothesis in this case is that nature, and thus causal inferences cannot be
specific factors do not account for the benefits drawn from their results. However, more recent
of psychotherapy, and that the null hypothesis studies on the alliance–outcome association
should not be rejected until sufficient evidence have tried to push the ball forward by disen-
in favor of the specific factors model is gath- tangling the temporal sequence of change. A
ered. He claims that the aggregate of treatment selection of them is mentioned here, as they
studies shows that all psychotherapies are give an impression of the direction the field is
equally effective, and that those studies that taking.
do show differences between psychotherapies Several studies show that (early) alliance
are probably the result of sampling error. In ratings predict (later) symptom change, con-
contrast, the evidence is consistent with models trolling for prior symptom improvement.
that emphasize common factors, according to Barber et al. (2000) found that alliance in
Wampold: alliance and outcome are strongly psychodynamic therapy predicted subsequent
related, and outcome variability due to thera- change in depression when prior change
pists is far greater than variability due to type in depression was partialed out. Alliance
of psychotherapy. His final conclusion is that also predicted outcome controlling for early
how therapy is conducted is more important change in couple’s therapy (Anker, Owen,
than what therapy is conducted. Duncan, & Sparks, 2010), but only in a
Kazdin (2005) finally argues that we need subsample of couples with four or more
new lines of research to resolve this debate. sessions. Crits-Christoph, Gibbons, Hamilton,
He too states that the presence or the absence Ring-Kurtz, and Gallop (2011) found that
of differences among types of psychotherapy session-to-session change in the alliance pre-
does not necessarily clarify the role of common dicted session-to-session change in symptoms
factors in therapeutic change. As Kazdin points in alliance-fostering treatment. Interestingly,
out, even when a certain type of psychotherapy they also found evidence for reversed causation
is found to be superior to another form, it (i.e., change in symptoms preceding change in
could be explained by the greater credibility of alliance) in later sessions. Although not proof
that treatment and the greater expectancies of of a causal link, these observational findings do
improvement it yields. Likewise, the evidence suggest that a change in alliance might account
on the association between therapeutic alliance for improvements in psychotherapy.
and outcome does not support the common In cognitive therapy (CT) for depression,
factor model. Kazdin argues that the way out however, studies from the DeRubeis research
of this deadlock is replicated, experimental group in Philadelphia show that alliance does
 COMMON (NONSPECIFIC) FACTORS IN PSYCHOTHERAPY 5

not predict outcome, or only to a small extent. Methodological Issues and Future
Feeley, DeRubeis, and Gelfand (1999) found Directions
that alliance did not predict change in symp-
toms. Prior symptom change did predict the There is no definitive answer to the question of
which factors—common, specific, or both—are
alliance, but only at a trend level. Strunk,
accountable for psychotherapeutic change. As
Brotman, and DeRubeis (2010) found that
Kazdin (2005) concludes, we need new lines
alliance did not predict subsequent symptom
of research to identify the active ingredients
change, but the other way around. In a sample
of psychotherapy: experiments, research into
of depressed patients who received both CT
therapist variations and individual change
and antidepressant medication, the alliance no
trajectories.
longer predicted session-to-session symptom
Research into mechanisms of change is
change when controlling for prior symptom
complex. The best approach to investigate
change (Strunk, Cooper, Ryan, Derubeis, &
causal pathways is to conduct experiments
Hollon, 2012). However, when looking at
in which common factors are isolated and
different components of the alliance, it was
directly manipulated. For therapy factors such
found that early session symptom change
as the therapeutic alliance, this is difficult (but
was associated with therapist–patient agree-
not impossible) to do. An alternative approach
ment of goals and tasks of therapy, and not
is to conduct controlled treatment studies
with the affective bond between therapist and
in which common factors, specific factors
patient (Webb et al., 2011). At a later stage and outcome are repeatedly measured during
in therapy, both therapist–patient agreement the entire course of therapy, to establish the
and affective bond were predicted by prior temporality of change (which process changes
symptom change, which points to reversed first?). Such a design can be used to conduct
causation. In another study, Webb et al. (2012) mediation analyses in which the temporality
investigated predictors of symptom change in aspect is accounted for. If several single medi-
two different patient samples who received ators are identified, multiple mediator models
CT, and found that the use of CT techniques can be built in which the relative contribution
was more strongly associated with subsequent of each mediator is investigated. Finally, it is
symptom change than alliance in one sample, possible to study the interaction of media-
whereas the reverse pattern was found in the tors (e.g., between therapeutic alliance and
other sample, suggesting that the nature of the therapeutic techniques) and differential med-
sample may moderate the effects of alliance itational processes in subgroups of patients
on outcome. Moreover, they found none of (moderated mediation). These models can be
the alliance by techniques interactions they advanced by taking therapist variation into
investigated to be significant, suggesting that account and investigating change trajectories
a strong alliance does not make patients more of individual patients. Obviously, this kind
responsive to therapeutic techniques. of research demands highly sophisticated
The evidence on the nature of the association designs and large sample sizes. For now, the
between alliance and outcome remains mixed. question of which factors are responsible for
However, one explanation for these differen- the benefits of psychotherapy remains an
tial findings is that in psychotherapies that empirical one.
place more emphasis on the alliance, such as
psychodynamic therapy and alliance-fostering
SEE ALSO: Designs for Studying Mediation;
therapy, alliance is more strongly linked to Dodo Bird Verdict in Psychotherapy; Mediation
subsequent symptom change than in psy- Analysis; Psychotherapy Process and Outcome
chotherapies that place less emphasis on the Research; Therapeutic/Working Alliance; Treatment
alliance, such as cognitive therapy. Adherence/Compliance
 6 COMMON (NONSPECIFIC) FACTORS IN PSYCHOTHERAPY

References Lambert, M. J. (1992). Psychotherapy outcome

research: Implications for integrative and eclectic
Anker, M. G., Owen, J., Duncan, B. L., & Sparks, J.
therapists. In J. C. Norcross & M. R. Goldfried
A. (2010). The alliance in couple therapy: Partner
(Eds.), Handbook of psychotherapy integration.
influence, early change, and alliance patterns in a
New York: Basic Books.
naturalistic sample. Journal of Consulting and
Lambert, M. J., & Ogles, B. M. (2004). The efficacy
Clinical Psychology, 78, 635–645.
and effectiveness of psychotherapy. In M. J.
Barber, J. P., Connolly, M. B., Crits-Christoph, P.,
Lambert (Ed.), Bergin and Garfield’s handbook of
Gladis, L., & Siqueland, L. (2000). Alliance
psychotherapy and behavior change. New York:
predicts patients’ outcome beyond in-treatment
John Wiley & Sons.
change in symptoms. Journal of Consulting and
Strunk, D. R., Brotman, M. A., & DeRubeis, R. J.
Clinical Psychology, 68, 1027–1032.
(2010). The process of change in cognitive
Castonguay, L. G., & Grosse Holtforth, M. (2005).
therapy for depression: predictors of early
Change in psychotherapy: a plea for no more
inter-session symptom gains. Behaviour Research
“nonspecific”and false dichotomies. Clinical
and Therapy, 48, 599–606.
Psychology: Science and Practice, 12, 198–201.
Strunk, D. R., Cooper, A. A., Ryan, E. T., Derubeis,
Craighead, W. E., Sheets, E. S., & Bjornsson, A. S.
R. J., & Hollon, S. D. (2012). The process of
(2005). Specificity and nonspecificity in
change in cognitive therapy for depression when
psychotherapy. Clinical Psychology: Science and
combined with antidepressant medication:
Practice, 12, 189–193.
Predictors of early intersession symptom gains.
Crits-Christoph, P., Gibbons, M. B., Hamilton, J.,
Journal of Consulting and Clinical Psychology, 80,
Ring-Kurtz, S., & Gallop, R. (2011). The
730–738.
dependability of alliance assessments: the
Tolin, D. F. (2010). Is cognitive-behavioral therapy
alliance-outcome correlation is larger than you
more effective than other therapies? A
might think. Journal of Consulting and Clinical
meta-analytic review. Clinical Psychology Review,
Psychology, 79, 267–278.
30, 710–720.
DeRubeis, R. J., Brotman, M. A., & Gibbons, C. J.
Wampold, B. E. (2005). Establishing specificity in
(2005). A conceptual and methodological
psychotherapy scientifically: design and evidence
analysis of the nonspecifics argument. Clinical
issues. Clinical Psychology: Science and Practice,
Psychology: Science and Practice, 12, 174–183.
12, 194–197.
Feeley, M., DeRubeis, R. J., & Gelfand, L. A. (1999).
Webb, C. A., DeRubeis, R. J., Amsterdam, J. D.,
The temporal relation of adherence and alliance
Shelton, R. C., Hollon, S. D., & Dimidjian, S.
to symptom change in cognitive therapy for
(2011). Two aspects of the therapeutic alliance:
depression. Journal of Consulting and Clinical
Differential relations with depressive symptom
Psychology, 67, 578–582.
change. Journal of Consulting and Clinical
Fluckiger, C., Del Re, A. C., Wampold, B. E.,
Psychology, 79, 279–283.
Symonds, D., & Horvath, A. O. (2012). How
Webb, C.A., DeRubeis, R. J., Dimidjian, S., Hollon,
central is the alliance in psychotherapy? A
S. D., Amsterdam, J. D., & Shelton, R. C. (2012).
multilevel longitudinal meta-analysis. Journal of
Predictors of patient cognitive therapy skills and
Counselling Psychology, 59, 10–17.
symptom change in two randomized clinical
Frank, J. D. (1961). Persuasion and healing: A
trials: the role of therapist adherence and the
comparative study of psychotherapy. New York:
therapeutic alliance. Journal of Consulting and
Schocken Books.
Clinical Psychology, 80, 373–381.
Gaston, L. (1990). The concept of the alliance and its
role in psychotherapy: theoretical and empirical
considerations. Psychotherapy, 27, 143–153.
Kazdin, A. E. (2005). Treatment outcomes, Further Reading
common factors, and continued neglect of Weinberger, J. (1995). Common factors aren’t so
mechanisms of change. Clinical Psychology: common: the common factors dilemma. Clinical
Science and Practice, 12, 184–188. Psychology: Science and Practice, 2, 45–69.

View publication stats