Facets of the psychotherapy relationship: a metaphorical approach.

Although many separate aspects of the psychotherapy relationship have been studied, including empathy, working alliance, and self-disclosure, a metaphorical approach has the potential to generate a more holistic perspective of this phenomenon. Hence, the goal of this study was to explore the nature of the psychotherapy relationship from the psychotherapists' perspective using a metaphorical approach. In an online survey, a sample of N=373 Czech psychotherapists and counselors rated a set of relational metaphors in terms of how accurately they depicted their roles in their relationships with their clients. The single most endorsed metaphor for the practitioner's role was a guide. Furthermore, the principal component analysis identified three relational components, namely, Mentor, Resource Supplier, and Remedy Distributor. The associations among these components and multiple practitioners' variables, including demographic and practice-related variables and theoretical orientation, were explored. These three components represent general dimensions of the psychotherapy relationship that cut across various theoretical orientations and, thus, define psychotherapy relationships in a general sense. ©Copyright: the Author(s).

Introduction

The psychotherapy relationship is considered a fundamental condition for and a means of psychotherapy change (Norcross, 2011). It is present in all psychotherapy systems, and it is probably the most often cited common factor in psychotherapy and counseling (Grencavage & Norcross, 1990). The psychotherapy relationship is a complex phenomenon that can be broadly defined as “the feelings and attitudes that counseling participants have toward one another, and the manner in which these are expressed” (Gelso & Carter, 1985, p. 159). Many partial aspects of the relationship have been recognized and studied empirically. In their tripartite model, Gelso and Carter distinguished three constituent parts of the psychotherapy relationship: working alliance, transference-countertransference configuration, and the real relationship. Of these three, the working alliance has been most extensively subjected to empirical research and has been found to be a robust predictor of outcome (Horvath, Del Re, Flückiger, & Symonds, 2011). Other aspects of the psychotherapy relationship that received empirical support include empathy, congruence, and unconditional positive regard (Crits-Christoph, Gibbons, & Mukherjee, 2013), group cohesion (Burlingame, McClendon, & Alonso, 2011), therapist self-disclosure (Ziv- Beiman, 2013), countertransference (Gelso & Hayes, 2007), and the reparation of alliance ruptures (Safran, Muran, & Eubanks-Carter, 2011).

Psychotherapists may differ in the relational manner with which they deliver the treatment (Fernández-Álvarez, García, Lo Bianco, & Corbella Santomá, 2003; Řiháček & Roubal, 2017a). Furthermore, the degree of match between a therapist’s relational style and a client’s relational preferences may be an important factor influencing treatment outcomes (Cooper, Norcross, Raymond-Barker, & Hogan, 2019). Therefore, it is essential to understand the variety of relational styles used by therapists in practice. Although the abovementioned concepts allow us to evaluate the quality of the therapeutic relationship and its connection to the outcome, they do not convey much information about how the relationship is represented in the therapists’ or clients’ immediate experience. To gain new insights into this problem, we applied a metaphorical approach.

Metaphorical approach

Well-defined concepts are crucial to understand reality and communicate this understanding to others. However, when concepts become used in a too rigid and definitive manner, they may eventually prevent further evolution of a field (Kuhn, 1996). Metaphors are fuzzy and experiential in nature. Unlike analytic definitions, they are rich in connotative meaning, convey implicit and emotional content, and are sensitive to culture-based experience (Lakoff & Johnson, 2003). Metaphors can make accessible meanings that are otherwise difficult to verbalize. Therefore, they may be particularly suitable to capture complex relational meanings beyond established concepts.

Metaphors are ubiquitous in our everyday lives. As Lakoff and Johnson (2003) put it: “Our ordinary conceptual system, in terms of which we both think and act, is fundamentally metaphorical in nature” (p. 4). They have been used in psychology from its very beginning and have often functioned as predecessors of formal models (Leary, 1990). The vital role of metaphors in all major psychotherapy schools has been advocated by Kopp (1995) and they have been plentifully used in psychotherapy research as well (e.g., alliance ruptures). Metaphors may thus help us consolidate a common-sense view of psychotherapy.

Relational metaphors in psychotherapy and medicine

The use of metaphors has been studied in the context of the psychotherapy process (Angus & Rennie, 1988, 1989; Rowat, Stefano, & Drapeau, 2008) and supervision (Young & Borders, 1999). Austin, Bergum, Nuttgens, & Peternelj-Taylor (2006) argued for metaphors as a suitable means for reflection of professional helping relationships. However, we are not aware of any study that has used a metaphorical approach to explore the nature of the psychotherapy relationship systematically. We have found only anecdotal attempts, such as that of Corsini (2008), who compared psychotherapists to detectives and counselors to teachers.

Since the therapeutic relationship shares some aspects with the doctor-patient relationship, relational metaphors developed in the field of medicine may possess some relevance for psychotherapy as well. Beisecker and Beisecker (1993) reflected on two leading metaphors of the physician-patient relationship, namely, paternalism and consumerism. Whereas paternalism denotes a relationship based on the physicians’ obligations, beneficence, trust, and principles of good medical care, consumerism stresses the patients’ rights, accountability, individual preferences, and values (Beisecker & Beisecker, 1993). Less formally, Carter (1989) explored other metaphors, such as parent, fighter, technician, teacher, friend, stranger, coach, repair person, or fellow detective, all of which have been used to describe various aspects or forms of the physician-patient relationship.

Aims of the study

Accordingly, this study aimed to explore the nature of the psychotherapy relationship from the perspective of the psychotherapist using a metaphorical approach. To address this aim, data from a national survey on Czech psychotherapists and counselors were used and subjected to the principal component analysis. Since this was an exploratory study, no a priori hypotheses were formulated about the components. However, we expected that the components would be related to the therapists’ gender, practice-related variables, and theoretical orientation.

Materials and Methods

Participants

A sample of N=373 Czech psychotherapists and counselors, 64% of whom were female, participated in the study. Their age varied between 25 and 71 years (M=40.17, SD=10.17), and their length of practice ranged from 0.5 to 45 years, with 62% of the participants having ten or fewer years of practice. While 51.7% had fully completed their training at the time of data collection, the remaining participants were in various stages of training completion. Regarding theoretical orientations, the most endorsed, as measured on a scale of 0 to 5, was the humanistic/ experiential orientation (M=3.08, SD=1.51), followed by the systemic/postmodern orientation (M=2.94, SD=1.65), the psychodynamic/psychoanalytic orientation (M=2.66, SD=1.61), and the cognitive/behavioural orientation (M=1.96, SD=1.19).

The participants represented various primary professions: psychology (62.5%), social work (19.0%), pedagogy (17.4%), psychiatry (5.9%), non-medical health professions (3.2%), addictology (2.4%), medicine other than psychiatry (1.9%), and others. As some participants stated more than one primary profession, the sum exceeds 100%. The participants practiced psychotherapy in various settings, including private practice (49.6%), social services (34.3%), health services (33.8%), educational settings (16.1%), and armed and rescue services (8.6%). Again, as some participants practiced in multiple settings, the sum exceeds 100%.

Instruments

A questionnaire was developed for this study. The first part contained demographic items, i.e., age, gender, and primary profession. The second part was composed of a list of relational metaphors that was developed to capture a range of aspects of the psychotherapy relationship. The list was developed in three steps. First, the authors (two professors with 15+ years of psychotherapeutic practice and a doctoral student) each brainstormed a list of metaphors for the therapists’ role which were intended to cover a broad range of possible meanings. Second, they merged the lists and removed metaphors that they deemed irrelevant and merged those that were similar in meaning. The reduction was motivated by the fact that the metaphor section was a part of a longer questionnaire and, therefore, had to be kept brief. Third, the authors critically inspected the list to see whether any essential aspect was missing. Furthermore, a draft version of the questionnaire was presented to 20 psychotherapists to gather their comments and suggestions. This process resulted in a list of 12 relational metaphors. The list was presented to the respondents with the following instruction: “Imagine that your relationship to your client could be expressed by a metaphor. Please, tell us to what extent does each of the following metaphors capture how you approach your clients. Take into account the last year of your psychotherapeutic/counseling practice.” Each metaphor was then rated on a six-point Likert-type scale, as follows: 0=never, 1=rarely, 2=sometimes, 3=often, 4=in most sessions, 5=in every session.

The third part of the questionnaire focused on trainingand practice-related items and included length of time practicing (in years); field of practice, such as health care, social services, educational setting, armed and rescue services, private practice, or other; highest level of client problem severity in one’s practice; workload, which was measured as an average number of sessions per week; theoretical orientation; frequency of supervision; and level of training completion. All practice-related questions were framed by an introductory statement that asked participants to answer the question based on the last year of their practice.

Client problem severity was measured by an item adopted from the Development of Psychotherapists Common Core Questionnaire (DPCCQ, Orlinsky & Rønnestad, 2005) and was composed of eight severity levels from “absent or minimal symptoms; socially effective; generally satisfied with life; no more than everyday problems or concerns” to “real danger of hurting self or others [e.g., suicide attempts, recurrent violence] or failure to maintain minimal personal hygiene, or gross impairment in communication [e.g., largely incoherent or mute].” Participants were asked to choose the highest severity level they commonly encountered in their practice. This score was used to represent the level of impairment of the participant’s clientele.

Broad-band theoretical orientation was measured using four items to investigate an endorsement of each of the four broadly defined theoretical orientations: cognitive/ behavioral, humanistic/experiential, psychodynamic/ psychoanalytic, and systemic/postmodern. The level of endorsement was measured on a six-point scale, as follows: 0=not at all, 1=slightly, 2=somewhat, 3=moderately, 4=greatly, 5=very greatly (adopted from the DPCCQ, Orlinsky & Rønnestad, 2005).

The questionnaire was part of a wide national survey and contained several other items that were not analyzed as part of this study. Partial results based on the same dataset were previously reported in Řiháček and Roubal (2017b, 2019, 2020).

Procedure

The project was approved by the Research Ethics Committee of the by Masaryk University (ref. no. EKV- 2016-006). A web-based version of the questionnaire was developed using the LimeSurvey platform (LimeSurvey Project Team, 2015). To maximize the sample size, several recruitment methods were used. First, umbrella associations for Czech psychotherapy and counseling practitioners were asked to distribute the questionnaire to their membership. Second, all Czech psychotherapy training institutes were asked to distribute the questionnaire to their graduates and their trainees. Third, personal e-mails and Facebook were used to increase participation.

Statistical analysis was conducted using R version 3.4.0 (R Core Team, 2017). To find a more parsimonious description of the set of relational metaphors, principal component analysis (PCA) was conducted (Henson & Roberts, 2006) on the metaphor items. Oblique rotation (direct oblimin) was chosen to rotate the components, as it was reasonable to expect correlated components. The number of components to extract was determined by the means of Horn’s parallel analysis (Horn, 1965) using the 95th percentile criterion (O’Connor, 2000), Kaiser’s criterion (eigenvalue >1) and a scree plot, as well as by the interpretability of the model. Finally, component scores were derived for each participant based on the structure matrix. PCA was conducted using the psych package (Revelle, 2016), and parallel analysis was conducted using the hornpa package (Huang, 2015).

To predict the use of the components identified in the previous step, a multiple regression analysis was conducted for each component. Two regression models were fitted. In Model 1, gender and four practice-related variables, namely, length of practice, highest level of severity, workload, and frequency of supervision, were used to predict component scores. In Model 2, four variables measuring practitioners’ endorsement of the four broadband theoretical orientations were added to the predictors used in Model 1. Predictors were entered into the model using the forced entry method and no centering was applied.

The homoscedasticity and linearity assumptions were checked visually, the absence of multicollinearity was assessed using the variance inflation factor, independence of errors was tested using the Durbin-Watson test, the distribution of residuals was tested using the Shapiro-Wilk test, and visual confirmation was conducted. The regression analysis was conducted using the car (Fox & Weisberg, 2011) and lmSupport (Curtin, 2017) packages.

Results

Descriptive analysis

We obtained 794 responses, 416 of which were incomplete and thus excluded[1]. Five respondents stated they had no psychotherapy training experience at all and, therefore, were excluded from the study. This yielded a final sample of N=373.

Descriptive statistics of the metaphor variables are reported in Table 1. The single most endorsed metaphor was guide, with a median value of 4 (in most sessions). It was followed by confessor, with a median value of 3 (often), and detective and consultant, with median values of 2 (sometimes). All other metaphors were perceived as rarely characterizing the respondents’ work with clients. The least fitting metaphor was educator, where most of the respondents reported they never adopted this role.

Table 1.

Description of metaphor variables.

Metaphor	Mdn	Interquartile range (25%; 75%)	M	SD	SEM
A doctor who treats an illness	1	0; 1	0.81	0.95	0.05
A guide who accompanies a client/patient in his or her personal development	4	3; 5	3.77	1.05	0.05
A coach who trains a client/patient	1	1; 2	1.50	1.15	0.06
A teacher who gives homework to a client/patient and who educates him or her	1	0; 2	1.14	0.96	0.05
An educator who oversees a client/patient	0	0; 1	0.62	0.88	0.05
A detective who strives to trace the source of a client/patient’s problems	2	1; 3	1.78	1.23	0.06
An adviser who offers recommendations	1	1; 2	1.57	1.03	0.05
A consultant who offers an independent assessment to a client/patient	2	1; 3	1.88	1.22	0.06
A confessor to whom a client/patient confides what lies heavy on him or her or what he or she is ashamed for	3	2; 3	2.54	1.25	0.06
A repairer who fixes what is not working	1	0; 1	0.77	0.97	0.05
A model for the client/patient	1	0; 2	1.23	1.10	0.06
A friend or caregiver who the client/patient does not have in the real world	1	1; 2	1.60	1.18	0.06

The metaphors appear in the same order as they did in the questionnaire.

Principal component analysis

The Kaiser-Meyer-Olkin measure verified the sampling adequacy for the analysis, KMO=.74, and the KMO values for individual items ranged from .55 to .85. Bartlett’s test of sphericity, χ (66)=970.97, P<0.001, indicated that correlations between items were sufficiently large for the PCA. While a parallel analysis indicated that three components should be extracted, Kaiser’s criterion and the scree plot converged on a four-component solution. We explored both solutions and found the three-component model to be superior in terms of interpretability and cleaner structure.

Table 2 presents the rotated component solution. The first component was defined by four metaphors, all of which share a mentor-type attitude, i.e., coach, teacher, educator, and adviser, and thus, this component was termed Mentor. The second component consisted of a set of five metaphors that embodied the aspect of an intimate, emotionally laden relationship in which the therapist personally invests himself/herself, i.e., confessor, friend, and model, or acted as a provider of information or opinion, i.e., detective and consultant. In either case, the therapist became a resource for the client to use either informationally or emotionally, and therefore, we named this component Resource Supplier. The third component was defined by two metaphors expressing an emphasis on pathology and expert attitude, i.e., doctor and repairer, and by the metaphor of educator, which tends to share the concept of expert attitude with the first two. Furthermore, it was negatively defined by the metaphor of guide, which tends to have, at least in the context of psychotherapy, the opposite connotations, i.e., non-expertise and a non-pathologizing attitude. Consequently, this dimension was termed Remedy Distributor. Together, these components explained 52.8% of the variance, and the component intercorrelations were practically negligible (r=0.19, r=0.12, and r=0.07).

Table 2.

Summary of principal component analysis of the relationships’ metaphors (pattern matrix).

Items	Mentor	Resource Supplier	Remedy Distributor	com.
Coach	0.81		–0.38	0.71
Teacher	0.81		0.12	0.71
Educator	0.62		0.40	0.60
Adviser	0.55	0.20	0.32	0.53
Confessor		0.80		0.62
Friend		0.75	–0.11	0.55
Detective		0.53	0.22	0.33
Model	0.32	0.50		0.41
Consultant	0.31	0.42		0.33
Guide		0.23	–0.76	0.61
Doctor		0.15	0.66	0.50
Repairer	0.22	0.34	0.44	0.43
Eigenvalue	3.30	1.64	1.39
% of variance	19.7	18.5	14.6

Component loadings over 0.40 appear in italics; loadings less than 0.10 were removed.

Prediction of component scores

Two multiple regression analyses (Model 1 and Model 2) were run for each component, and the results are presented in Table 3. All predictors were dichotomized using the median as the split point because they were either ordinal or non-normally distributed. Three observations were excluded due to missing data for one or more predictors. No multivariate outliers were removed. However, as the distribution of the predicted variable (i.e., component scores) was positively skewed in two cases, it was transformed using exponential transformation to produce normally distributed residuals. Furthermore, three univariate outliers were removed in the case of the third component for the same reason.

Table 3.

Results of regression analyses.

Predictors	Mentor		Resource Supplier		Remedy Distributor
	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
Gender°	–0.05	–0.07	0.00	0.03	0.04	0.04
Length of practice	–0.02	–0.01	0.06	0.03	0.21***	0.19***
Highest problem severity	–0.01	–0.04	–0.04	–0.05	0.15**	0.13*
Workload	–0.10	–0.08	0.13*	0.07	–0.03	–0.02
Frequency of supervision	–0.04	–0.03	0.03	–0.01	–0.01	–0.02
Theoretical orientation
Cognitive/behavioral		0.39***		0.15**		0.10
Humanistic/experiential		0.00		0.25***		–0.17**
Psychodynamic/psychoanalytic		–0.12*		0.16**		0.09
Systemic/postmodern		–0.09		–0.13*		–0.02
R2	0.02	0.21***	0.03	0.19***	0.07***	0.11***
ΔR2		0.19***		0.16***		0.04**
n	370		370		367

N=370. Intercepts were omitted. Values represent standardized regression coefficients. Asterisks indicate significance level: *P<0.05, **P<0.01, and ***P<0.001. °A positive value means that a component was preferred by women.

Model 1 was nonsignificant for the Mentor component. In Model 2, however, the Mentor component statistically significantly predicted by the cognitive/behavioural orientation (β=0.39, P<0.001) and negatively predicted by the psychodynamic/psychoanalytic orientation (β= –0.12, P<0.05).

Although one variable, i.e., workload, reached significance in Model 1 for the Resource Supplier component, the overall model was nonsignificant. In Model 2, this component was positively predicted by the humanistic/experiential orientation (β=0.25, P<0.001), psychodynamic/ psychoanalytic orientation (β=0.16, P<0.01), and cognitive/behavioural orientation (β=0.15, P<0.01) and negatively predicted by the systemic/postmodern orientation (β= –0.13, P<0.05).

Remedy Distributor was significantly predicted by length of practice (β=0.21, P<0.001) and highest level of severity (β=0.15, P<0.01) in Model 1. In Model 2, both variables remained significant. Furthermore, the component was negatively predicted by the humanistic/experiential orientation (β= –0.17, P<0.01).

Discussion

Using data from an online survey on Czech psychotherapists and counselors, we explored facets of psychotherapy relationships from the practitioner’s perspective expressed using a set of metaphors. The descriptive analysis revealed that the single most widely endorsed metaphor for the practitioner’s role in the therapeutic relationship was guide. This metaphor signifies a non-directive companion who helps clients find their way to a desired state of being, which is consistent with the Rogerian concept of the therapeutic relationship (Rogers, 1957). Today, Rogers’ contribution is considered as a pan-orientation concept (Cooper, 2008) that likely resonates with most practitioners’ experiences. However, the reason for the popularity of this metaphor may also be that it is most likely used by psychotherapy trainers to define the role of the therapist, and it may have been simply reproduced by the participants as a cliché, without giving it the same amount of critical consideration as they did the other metaphors. Nevertheless, except for educator, all other metaphors were perceived to capture some aspect of the psychotherapist’s role as well.

Principal component analysis indicated the existence of three underlying components of the metaphors, specifically, Mentor, Resource Supplier, and Remedy Distributor. The Mentor component was defined by metaphors that suggested the practitioner teaches, trains, or supervises the client. In this sense, they share an element of directivity and explicit guidance or mentoring. However, they also seem to share an implicit non-pathologizing attitude, a focus on clients’ personal growth, and the utilization of their potential. From the perspective of a teacher or a coach, a client is not somebody to be fixed or cured. Rather, it is someone in need of direction, instruction, and supervision. Furthermore, having a lack of skills is not an illness per se. Instead, it is a natural developmental stage from which a client can evolve to another, more advanced stage.

The Mentor component was more highly endorsed by practitioners with a cognitive/behavioral orientation, which agrees with the more directive nature of this orientation. Furthermore, it was negatively related to the psychodynamic/ psychoanalytic orientation, which also corresponds to the relative reluctance of insight-oriented therapists to provide their clients with explicit instructions (Scaturo, 2002). Surprisingly, the Mentor component was unrelated to any of the demographic and practice-related variables, suggesting that it is a matter of a therapist’s personal preference or, more likely, client-treatment matching (Beutler & Harwood, 2000).

The Resource Supplier component was composed of a group of metaphors that suggest the therapist is available as a resource for the client. The therapist may offer many different resources, including information or counsel, new insights or perspectives, a close relationship, a safe space for the client’s disclosure, empathy, support, acceptance, inspiration, or even a personal example for the client to follow. The influence is, however, non-directive in that the therapist does not impose his or her influence on the client, and the client is free to choose how to use the resources being offered. This component resembles Beisecker and Beisecker’s (1993) metaphor of consumerism.

The reported use of the Resource Supplier component was positively associated with the humanistic/experiential, psychodynamic/psychoanalytic, and cognitive/behavioral orientations. The more practitioners endorsed any of these orientations, the more they tended to identify the role of a resource provider as descriptive of their therapeutic approach. This makes sense if we consider the orientations as reservoirs of the resources, whether they be psychodynamic interpretations, authentic relationships, provision of corrective experience, or the therapist modeling of assertive behavior. Surprisingly, the use of this component was negatively related to the systemic/postmodern orientation, which suggests that practitioners of this orientation are less inclined to reflect on their work this way and that, instead, they strive to adopt a more neutral attitude (Schlippe & Schweitzer, 1998). Furthermore, practitioners who indicated higher workloads reported that they tended to find this component slightly more descriptive of their approach than those practitioners who reported fewer sessions per week, although this was partly confounded with theoretical orientation. This is an interesting finding with no straightforward interpretation. It is possible, for instance, that those who have a more intensive practice do so because they are more comfortable providing many types of resources to their clients, or because, based on their more intensive immersion in psychotherapy practice, they consider themselves more skilled at providing these resources.

The metaphors loading on the third component, Remedy Distributor, characterize a perspective from which clients are ill or disordered and therapists are experts who can cure them or repair what has gone wrong. It is implicit in this perspective that the clients do not possess the resources or the abilities needed to recover and that therapist expertise is necessary to produce change. Consistent with the medical model, the therapist is the agent of change, and the client is a recipient of the treatment. The three metaphors that loaded on this component received the lowest average scores which means that, overall, the participants found them rather unsuitable to describe the role of the therapists. This corresponds to the fact that the medical model has been largely criticized as inconsistent with psychotherapy (Elkins, 2009; Wampold, 2001). However, this was not true for all participants and the existence of this component suggests that the medical model may be relevant under some circumstances. This component can be roughly compared to Beisecker and Beisecker’s (1993) metaphor of paternalism.

It is further evident that this component was endorsed more by practitioners who worked with a more disturbed clientele, which suggests that the Remedy Distributor role is more easily evoked in this context. Furthermore, the endorsement of this component was negatively related to the humanistic/experiential orientation, which is consistent with the non-pathologizing nature of this orientation (Cain, 2002). Somewhat surprisingly, however, the endorsement was also higher with higher levels of practitioner experience. This relationship may have been confounded with the setting within which the practitioners worked: practitioners in health services and armed and rescue services reported, on average, twice as many years of practice than those in social services and educational settings. At the same time, respondents in the former two settings had clients with more severe problems than those in the latter two settings.

Limitations

Several limitations must be considered when generalizing the results of this study. First, the list of metaphors was non-exhaustive. It was composed by researchers with a humanistic/experiential orientation and even though a considerable effort was dedicated to its development, it may be biased. Future studies may employ a more extensive list of metaphors.

Second, the questionnaire captured an overall selfassessment of the suitability of the metaphors. It is possible that the meanings ascribed by participants to the metaphors differed from those assumed by the researchers. Moreover, the meanings of the metaphors may also not be easily transferable to other cultures and/or languages.

Third, the use of metaphors allows researchers to capture implicit and otherwise inaccessible meanings. A considerable amount of subjective inference and imagination was, however, required during the interpretation of the PCA results. Although the emphasis was placed on discussion and consensus, it is possible that another group of researchers would have arrived at a different interpretation of the components.

Fourth, the questionnaire was based on the participants’ self-assessment of their therapeutic approach. Although Castonguay et al. (2017) demonstrated that practitioners can recall accurately the types of techniques they used, Santa Ana et al. (2008) suggested that the reliability of self-report can be limited.

Fifth, the combination of several recruitment methods does not allow us to estimate the response rate. However, trying to impose more control over the recruitment process would significantly reduce the sample size. A considerable limitation is the low degree of representation of the cognitive/behavioral orientation relative to the remaining orientations. That said, this reflects that only two of more than 40 Czech training institutes teach cognitive-behavioral therapy.

Conclusions and implications

This study has indicated that the practitioner’s role in the therapeutic relationship can be explored using a metaphorical approach. The metaphor considered most accurate by the practitioners to capture the nature of the therapeutic relationship was that of guide. Furthermore, the analysis demonstrated that the plenitude of metaphors can be meaningfully reduced to three mutually independent components: a directive, mentor-like approach focused on personal growth (i.e., Mentor); a non-directive approach based on the provision of cognitive and/or emotional resources for the client (i.e., Resource Supplier); and an expert approach in which the client is perceived as a recipient of treatment (i.e., Remedy Distributor). The components were related to practitioners’ self-rated theoretical orientations in a non-trivial pattern, suggesting that these components capture dimensions cutting across these orientations and, thus, define psychotherapy and counseling relationships in a more general sense. In case the model is confirmed by future research, it will provide researchers, as well as practitioners, with a useful tool to reduce the complexity of the therapeutic role into a set of three independent components that may be related to different aspects of the psychotherapy process and outcome. Future studies can explore if these three aspects of the therapeutic relationship are differentially effective with various clients’ conditions (e.g., depression), characteristics (e.g., level of psychological mindedness), and expectations (e.g., supportive vs. confrontational approach).

Due to its imaginative and holistic character, the metaphorical approach can be used in psychotherapy practice and supervision. The three components, as well as the individual metaphors, can be used to reflect on the nature and quality of the therapeutic relationship with a client. Knowing what relational role a client attributes to the therapist and what he or she expects from the therapist in the therapeutic relationship is essential for the therapist to plan interventions. Metaphors can allow the therapist and the client to explore aspects of the therapeutic relationship that are inaccessible by the rational-analytic approach. However, metaphors should be used with caution. In a study on public political communication, Keefer and Landau (2015) have shown that relationship metaphors may elicit defensive information processing in people with avoidant attachment. Although we cannot automatically extrapolate this finding to the therapeutic context, therapists should exercise caution when using such metaphors in conversations with their clients.

Metaphors can be also used to reflect on the nature and form of the therapeutic relationship in the context of supervision and training. Trainees and supervisors may elicit metaphors that express their perception of a particular therapeutic case. Again, metaphors may help to verbalize the implicit and intuitive knowledge. Thus, they may sensitize trainees to various nuanced aspects of the psychotherapy relationship. However, readers should bear in mind that metaphors are not meant to replace rationalanalytic concepts and, thus, sacrifice the precision and unambiguity in the study of the therapeutic relationship. Albeit rich in connotative meaning, metaphors are fuzzy concepts that allow for multiple interpretations. Therefore, they are meant to serve as a complementary source of information that makes the picture more complete.