34
Re-visioning Empathy: Theory, Research and Practice
Jeanne C. Watson
Ontario Institute for Studies in
Education / University of Toronto
October 2000
Almost 60 years of research consistently demonstrates that therapist empathy is the most potent predictor of client progress in therapy. Yet many therapists continue to fail to appreciate its power, understand how it leads to change or how to respond empathically with optimum impact. In contrast this paper will present a view of empathy as an active ingredient of change that facilitates clients’ meta-cognitive processes and emotional self-regulation. A review of the research on empathy reveals that it is an essential component of successful therapy in every therapeutic modality (Burns & Nolen-Hoeksma, 1991; Raue & Goldfried, 1994; Greenberg et al, 1993; Rowe, 1997; Linehan, 1997; Luborsky et al, 1988; Rounsaville, Weissman & Prusoff, 1981; Strupp et al, 1969; Safran & Wallner, 1991; Rogers, 1959: 1975; Miller et al, 1980).
Empathy is a basic component of emotional intelligence (Goleman, 1995). It is a multidimensional and complex epistemological process and is a way of knowing that involves both our affective and cognitive systems. (Bohart & Greenberg, 1997; Duan & Hill, 1996; Feschbach, 1997). According to developmental theorists it is a capacity present from birth which develops over time as our emotional and cognitive systems mature. Titchener translated the term empathy from the German word “einfuhlung”, as "to feel one's way into" (Bozarth, 1997). It was originally used in German aesthetics as a means of coming to know a particular work of art. Later, empathy was used by Rogers (1959; 1975) and Kohut (1971; 1977) to describe a way of being towards others to promote healing in psychotherapy.
Numerous theorists see empathy as a basic relationship skill (Bohart & Greenberg, 1997; Davis, 1979; Hogan, 1969; Kohut, 1971; Rogers, 1975; Watson, Goldman & Vanaerschot, 1998). For interpersonal communication to occur a certain amount of empathy is required in order to understand others at even the most basic level (Feschbach, 1997; Hoffman, 1982; Jordan, 1997; Kohut, 1970; Linehan, 1997; Rogers, 1975; Trop & Stolorow, 1997). However, there are different levels of understanding and different types of empathic process. We can understand another by knowing what they mean intellectually, by comprehending their values, worldviews, goals and objectives, but to be truly empathic this understanding needs to be informed by knowing or understanding how things affect them emotionally. People's emotions reveal the significance or meaning of events for them (Taylor, 1975; Orlinsky & Howard, 1986; Greenberg, Rice & Elliott, 1993; Goleman, 1995; Rogers, 1965). Thus in order to understand and know another person it is important to be emotionally responsive to them.
The primary purpose of this chapter is to examine the role and function of empathy in psychotherapy. First, research on empathy from 1940 to the present will be examined. This will be followed by a discussion of how empathy informs humanistic/experiential therapists' practice. Finally a theoretical model of empathy's role in promoting change in psychotherapy will be presented.
Rogers' Three Therapist Conditions
For Rogers empathy was one of three essential conditions offered by therapists to their clients in psychotherapy. The other two therapist conditions were congruence and unconditional positive regard or a non-judgmental stance towards the other. The latter attribute was seen as vital to the provision of empathy. Rogers did not think it was possible to empathize fully and completely with others' experiences if one was evaluating them negatively. Moreover, as developmental theorists have observed, it seems highly unlikely that someone could be truly empathic if they feel hostility or anger towards the other person (Feschbach, 1997; Hogan, 1982). Thus in order for empathy to be therapeutic it needs to occur in a sympathetic, safe, and non-judgmental climate, where clients feel prized. In addition it is important for therapists to appear congruent if they are to be experienced as empathic.
Rogers (1965) defined empathy as the ability to perceive accurately the internal frames of reference of others in terms of their meanings and emotional components. This definition highlights empathy both as an emotional and a cognitive process. Rogers saw empathy as the ability to see the world through others' eyes so as to sense their hurt and pain and to perceive the source of their feelings in the same way as they do. In this regard he was careful to distinguish identification from empathy. Identification indicates a loss of boundaries as one adopts others' views of reality.
One way Rogers suggested that therapists could demonstrate empathy was through trying to reflect clients' feelings. However, this activity was often misconstrued as merely parroting clients' words or repeating the last thing they said (Rogers, 1975). This view reflects a misunderstanding of empathic responding, which is a sophisticated and highly complex way of being with clients. The resources required to listen empathically, to get at the heart of clients' communications, to distil the essence and centrality of their messages, is a taxing and demanding exercise. Research studies have shown that therapeutic empathy is correlated with therapists' cognitive complexity. Listening attentively to the nuances of clients’ narratives is a sophisticated exercise of critical deconstruction in the moment that requires great concentration.
The simplistic way in which empathy came to be understood as reflection of feelings was problematic and deterred Rogers from discussing it further until his landmark article in 1975. In his 1975 paper, Rogers returned to the subject of empathy and tried to define it more precisely. At that time Rogers saw empathy as a process, not a state of being. Thus it is an ongoing attempt to enter the private, subjective world of the other, while at the same time being sensitive to the changes in meaning so that they can be tracked accurately. The process of trying to understand the subjective world of the other is done with careful attention to nuance in order to sense meanings of which the person may not be fully aware. However, this does not mean uncovering or reporting on feelings and sensations of which the person has no awareness. Moreover, attempts to understand what clients are saying are always done tentatively and constantly checked with them. Clients remain the final arbiters of whether therapists have understood them correctly or caught a meaning of which they were not fully aware. Rogers (1975) saw empathy as a way to evoke self-directed change and to empower the person. Thus he took pains not to appear the expert with clients in therapy so that they would assume responsibility for their own change processes and behaviors, and come to see themselves as the best judge of their own needs and feelings.
Research on the Role of Empathy in Psychotherapy
The role of empathy in facilitating change received considerable attention in the research literature after Rogers posited that it, together with unconditional positive regard and congruence, were the necessary and sufficient therapist conditions of psychotherapeutic change. The findings generated considerable debate and heated discussion with respect to the efficacy of these three conditions with numerous researchers either supporting or discrediting Rogers’ position. After the initial flurry of interest that seemed to reveal conflicting findings, research on empathy and the other two conditions declined (Duan & Hill, 1996; Sexton & Whiston, 1994; Orlinsky, Grawe & Parks, 1994). The decline in empathy research has been attributed to a number of factors. First, difficulties defining the construct, second poor research tools (Barkham & Shapiro, 1986; Duan & Hill, 1996; Lambert De Julio & Stein, 1978; Patterson, 1983; Sexton & Whiston, 1994) and third, increased interest in the working alliance (Duan & Hill, 1996; Orlinsky, Grawe & Parks, 1994). In this section we will discuss how the concept of empathy has been defined and measured and evaluate its role in treatment.
An examination of the research on the role of empathy in psychotherapy reveals that it is an area plagued by methodological difficulties exposing the tensions in various researchers paradigmatic assumptions. A thorough review of the area suggests that researchers are, at times, comparing apples and oranges, and highlights the inherent difficulty they face in trying to measure a dynamic and interpersonal phenomenon. After Rogers' initial hypothesis the two most commonly used measures that were developed to measure empathy in the session were the Accurate Empathy Scale (Truax & Carkhuff, 1967) and the Relationship Inventory (Barrett-Lennard, 1962).
Measures of empathy
The Accurate Empathy Scale (Truax & Carkhuff, 1967) evaluates therapist responses, on a 10-point scale, in terms of the degree to which they communicate empathy. It is based on independent raters' evaluations of the therapeutic process. According to Duan and Hill (1996) it is one of the most common ways of measuring therapist empathy from an observer's perspective. The other way of measuring therapist empathy has been from the clients' perspective using Barrett-Lennard's Relationship Inventory (Barrett-Lennard, 1962). This is a self -report measure that asks clients to comment on whether they experienced their therapists as genuine, prizing and empathic. Not surprisingly, from our post-modern perspective, raters and clients provide different views and judgements of empathy. Correlations between external judges ratings and clients’ ratings of empathy are low (Bozarth & Grace, 1970; Kiesler, Mathieu & Klein, 1967; Rogers et al, 1967) or non-existent (Caracena & Vicory, 1969; Carkhuff & Burstein, 1970; Fish, 1970; Hill, 1974; Kurtz & Grummon, 1972; McWhirter, 1973; Truax, 1966; Van der Veen, 1970).
A number of reasons have been posited to account for these anomalous findings. The most cogent is that each measure uses a different criterion for determining empathy. According to this view it is likely that external raters and clients are attending to different cues on which to base their evaluations of therapeutic empathy. Moreover most raters are limited to audio recordings and thus are not able to apprise themselves of therapists' non-verbal behaviors as a means of communicating empathy (Duan & Hill 1996; McWhirter, 1973).
As noted by numerous reviewers it is ironic that despite Rogers’ emphasis on communicated empathy or clients' perceptions of their therapists empathy as the important determinant of outcome, researchers continue to use external evaluations (Barkham & Shapiro, 1986; Caracena & Vicory, 1969; Duan & Hill, 1986; Gurman, 1977; Orlinsky & Howard, 1976). The questionable validity of using observer's ratings is further underscored by the finding that ratings of therapist empathy can be made independently of client responses (Orlinsky & Howard, 1986). The rationale for not using clients as raters of therapeutic empathy was that they were likely to be unreliable and inclined to distort or incorrectly perceive their therapists' behavior (Gurman, 1977). This conclusion reveals a clash between certain research assumptions based in a positivist research paradigm that undervalues the participation of research subjects, and views of pathology that are at odds with the underlying assumptions of client-centered therapy (Bohart & Tallman, 1997; Watson & Rennie, 1994).
The Relationship between Empathy and Outcome
Numerous studies that have been conducted to determine whether empathy is indeed a necessary and sufficient condition along with the other therapeutic conditions of genuineness and unconditional positive regard. Overall, there is a preponderance of evidence to suggest that therapist empathy is indeed a crucial variable in psychotherapy, notwithstanding those studies that have failed to find significant relationships (Bergin, 1966; Gurman, 1977; Patterson, 1983; Luborsky, Crits-Christoph, Mintz & Auerbach, 1988; Orlinsky, Grawe & Parks, 1994). In their most recent review of psychotherapy process and outcome research Orlinsky et al (1994) note that, in the period 1972-1989, 54% of studies using either externally rated or client judged ratings of empathy support the relationship between therapists' communicating empathically with their clients and therapy outcome. Lambert, De Julio & Stein (1978) and Luborsky et al (1988) reviewed a total of 23 studies looking at the relationship between therapist empathy as rated by an external rater and therapy outcome over an 11 year period from 1962-1973. Of this total, 14 studies reported positive significant relationships between therapist empathy and outcome while 9 were not significant.
In their meta-analysis of the studies that they reviewed, Luborsky et al (1988) reported a mean r = .26. Lamb (1981), similarly, reported a weighted mean r = .26 in his meta-analysis of research studies which examined the efficacy of the therapeutic conditions of empathy, genuineness and positive regard. This is the same as that reported in a meta-analysis of the relationship between the therapeutic alliance and outcome conducted by Horvath & Symonds (1991). While the overall correlation of r = .26 is relatively small, it provides considerable support for the relationship between the presence of empathy and outcome, as well as that of the other therapeutic conditions and the quality of the therapeutic relationship.
Stronger evidence to support the hypothesis that therapists' empathy is related to client change in psychotherapy comes from studies that have examined the relationship between client perceived ratings of therapeutic empathy and outcome (Gladstein, 1977; Gurman, 1977; Orlinsky, Grawe & Parks, 1994). Most of these studies have employed the Barrett-Lennard Relationship Inventory to assess clients' experience of the relationship. Gurman (1977) in his review of studies conducted between 1954-1974 noted that 17 looked at the relationship between empathy and outcome. Of these, 14 reported a positive relationship between empathy and outcome and 3 found no relationship. Gurman (1977) notes that two of these three studies were analogue studies, and in one, treatment consisted of two sessions only. This is hardly enough time for clients to begin to have an adequate sense of whether their therapists accurately understood them or not. Barrett-Lennard suggested that ratings of empathy be done after the fifth session (Gurman, 1977). Other studies have shown that clients' perceptions of empathy as well as therapists' understanding of their clients vary over time (Cartwright & Lerner, 1965; Kalfas, 1974 in Orlinsky &Howard 1986; Kurtz & Grummon, 1972; Marangoni, Garcia, Ickes & Teng, 1995; Patterson, 1983).
A review of studies from 1976-1994 (Luborsky et al, 1988; Orlinsky et al; 1994) provides additional support for the relationship between therapists' empathy and outcome, with 8 of 10 studies reporting positive findings and 2 reporting no relationship. Studies that have conducted in-depth interviews with clients to determine the effective elements of treatment have consistently found that an important factor is the opportunity to talk with an understanding, warm, and involved person (Cross, Sheehan & Khan, 1982; Feiffel & Eels, 1963; Lietaer, 1990; Strupp, Fox & *Lesser, 1969; Watson & Rennie, 1994). Orlinsky et al (1994) note in their review that clients are the more discriminating than external raters and therapists with respect to therapist process variables. Moreover some studies have shown that therapists are not as discriminating of their own behavior and sometimes rate themselves as higher in empathy than clients or external raters (Gurman, 1977; Gross et al 1977; Kurtz & Grummon, 1972; Lafferty, Beutler & Crago, 1989).