The role of therapist self-disclosure in psychotherapy: A qualitative review Jennifer R. Henretty ���, Heidi M. Levitt 1 University of Memphis, Department of Psychology, 202 Psychology Building, Memphis, Tennessee, 38152-6400, United States a b s t r a c t a r t i c l e i n f o Article history: Received 18 February 2009 Received in revised form 7 September 2009 Accepted 14 September 2009 Keywords: Self-disclosure Psychotherapy process Psychotherapy guidelines Over 90% of therapists self-disclose to clients (Mathews, 1989 Pope, Tabachnick, & Keith-Spiegel, 1987 Edwards & Murdock, 1994), however, the implications of therapist self-disclosure are unclear, with highly divergent results from one study to the next. The goal of this paper was to review the empirical literature relevant to therapist self-disclosure, and provide the reader with a comprehensive understanding of the factors that affect, and are affected by, therapist self-disclosure. Findings are organized into an integrated model examining the who, what, when, why, and how of therapist self-disclosure. In addition, training implications and suggestions for future research are provided. �� 2009 Elsevier Ltd. All rights reserved. Contents 1. Review of the quantitative research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 1.1. Independent variables in the therapist self-disclosure research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 1.1.1. Independent variables related to the therapist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 1.1.2. Independent variables related to the client and/or therapeutic process . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 1.1.3. Independent variables related to issues of diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 1.2. Dependent variables in the therapist self-disclosure research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 1.2.1. Dependent variables related to the therapist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 1.2.2. Dependent variables related to the client and/or the therapeutic process . . . . . . . . . . . . . . . . . . . . . . . . . . 66 1.2.3. Dependent variables related to client self-disclosure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 1.3. Types of therapist self-disclosure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 1.4. Self-disclosure versus nondisclosure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 2. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 2.1. Research implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 2.2. Training implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 2.3. Clinical implications: Synthesizing the research and the theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 2.3.1. Nondisclosure is no longer the easy answer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 2.3.2. Therapists need to consider the issue now . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 AlthoughSidney Jourardwas not creditedwithbeing the first to use the term ���self-disclosure��� until 1958 (Gallucci, 2002), the debate about its use in psychotherapy had been ongoing for decades. Sigmund Freud conceived the first model of the ideal therapist stance, likening the therapist to a blank screen. ���The [therapist] should be opaque to his patients and, like a mirror, should show them nothing but what is shown to him��� (Freud, 1912/1958, p. 118). Thus, traditional psycho- analysts believed their goal as therapists was to remain a neutral and anonymous medium, upon which patients could project transference distortions for the purpose of interpretation. Conversely, the prodisclosure argument was developed within a humanistic framework. In the late 1950s, Rogerians were the first clinicians to adopt practices of therapist self-disclosure (Farber, 2006). Since then, client-centered therapists continually have argued Clinical Psychology Review 30 (2010) 63���77 ��� Corresponding author. E-mail addresses: jenhenretty@hotmail.com (J.R. Henretty), h.levitt@mail.psyc.memphis.edu (H.M. Levitt). 1 Tel.: +1 901 678 5489 fax: +1 901 678 4518. 0272-7358/$ ��� see front matter �� 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.cpr.2009.09.004 Contents lists available at ScienceDirect Clinical Psychology Review

that by cautiously modeling openness, strength, vulnerability, and the sharing of intense feelings, the therapist who uses therapy-relevant self-disclosure invites the client to follow the lead and cultivates trust, perceived similarity, credibility, and empathic understanding (Kottler, 2003 Knox, Hess, Petersen, & Hill, 1997). Furthermore, feminist and multicultural therapeutic approaches have since placed a premium on therapist self-disclosure (Brown & Walker, 1990 Mahalik, Van Ormer, & Simi, 2000). While the zeitgeist ever-changes and the debate continues, we know that therapists are self-disclosing. Although therapist self- disclosure is one of the rarest techniques, comprising an estimated average of 3.5% of therapist interventions (Hill & Knox, 2002), over 90% of therapists report that they have self-disclosed in therapy (Edwards & Murdock, 1994 Mathews, 1989 Pope, Tabachnick, & Keith-Spiegel, 1987). However, contradictory empirical findings and theoretical conceptualizations, teamed with ambiguous ethical guidelines (Domenici, 2006) and little to no training as to the nature and use of therapist self-disclosure (Beutler, Crago, & Arizmendi, 1986), may leave therapists feeling vulnerable and anxious about self-disclosing (Hill & Knox, 2002 Knox & Hill, 2003). Practitioners need a more coherent understanding of therapist self-disclosure so that they can weigh the pros and cons of its use, and, when deemed therapeutically appropriate, use this technique with confidence. Therefore, this paper aims to review the quantitative research pertaining to therapist self- disclosure and, in the Discussion section, synthesize those findings with the relevant theoretical literature into an integrated model examining the who, what, when, why, and how of therapist self- disclosure. In addition, training implications and concrete suggestions on how to address gaps and inconsistencies in the literature with future research will be provided. 1. Review of the quantitative research A perusal of the relevant literature suggests that, despite the many writings supporting judicious therapist self-disclosure, some authors remain concerned with clinical ramifications of such disclosure that is, they question whether therapist disclosure is a therapeutically helpful intervention(Peterson, 2002). This paperreviewsquantitative results of published research on verbal therapist self-disclosure via summaries and, when the research body consists of five or more studies, via tables. Note that every study included in this review is examined in terms of variables (e.g., client perceptions, therapist orientation) and types of comparisons (e.g., disclosure vs. nondisclosure, positive self-disclosure vs. negative self-disclosure) that it utilizes, thus some studies are necessarily referenced in multiple summaries and tables. The empirical reviewwillbe divided intothefollowingsections: (a)studiesexamining the relationship between various independent variables and therapist self-disclosure, (b) studies examining the relationship between various dependentvariablesandtherapistself-disclosure,(c)studiescomparing types of therapist self-disclosure, and (d) studies comparing disclosure to nondisclosure. 1.1. Independent variables in the therapist self-disclosure research This section is subdivided into independent variables associated with the therapist, which tend to be demographic (e.g., education, theoretical orientation, professional status, age), and independent variables associated with the client (such as client demographic or diagnostic variables) or with the therapeutic process (i.e., alliance, as it affects disclosure). A third subsection discusses independent variables associated with issues of diversity, consisting of gender, ethnicity, and sexual orientation. 1.1.1. Independent variables related to the therapist When examining the relationship between frequency, or use, of therapist self-disclosure and therapist-related independent variables, this review found no correlation with therapists' age (Berg-Cross, 1984) or with therapists' level of education (Andersen & Anderson, 1989 Simone, McCarthy, & Skay, 1998). Studies that attempted to discern, using analogue methodology (i.e., using non-client samples in situations designed to be similar to the therapy experience), whether and how therapists' professional status interacts with therapist self-disclosure on clients' perception and responses, had inconsistent results: One study foundthatprofessionalstatushadnomainorinteractioneffectonclients responses and perceptions (McCarthy, 1982) one study's findings implied that it may be more importantfor professional therapists to self- disclose than non-professional therapists (Merluzzi, Banikiotes, & Missbach, 1978) and one study's finding implied that paraprofessional therapists should self-disclose intimately, whereas professional thera- pists should utilize non-intimate self-disclosures (Simonson & Bahr, 1974). Because each study examined different types of therapist self- disclosure and only one used a no disclosure control, results cannot meaningfully be aggregated across studies. When therelationship between therapist self-disclosure andamount of clinical experience is examined (see Table 1), results across studies were inconsistent. However, it seems that early-career therapists may disclose less and for different reasons (e.g., to prevent client from questioning therapist's mental health Simone et al., 1998) than more experienced therapists, butthatthefrequency of self-disclosure plateaus at some point for female therapists, showing little difference between therapists with an intermediate amount of experience and more. For male therapists, self-disclosure may decrease as experience is gained past an intermediate amount. Therapists of differing theoretical orientations have different views on the utility and appropriateness of therapist self-disclosure (Peterson, 2002 Williams & Levitt, 2008). However, results of surveys that examined use of therapist self-disclosure and theoretical orientation (i.e., Andersen & Anderson, 1989 Berg-Cross, 1984 Edwards & Murdock, 1994 Simi & Mahalik, 1997) did not consistently show differences among the various orientations. Two surveys reported no relationship (Andersen & Anderson, 1989 Berg-Cross, 1984), whereas one survey (Edwards & Murdock, 1994) reported that psychoanalytic/ psychodynamic therapists utilized less therapist self-disclosure than Table 1 Does clinical experience affect amount of therapist self-disclosure (TSD)? (6). Possible answers Studies Summary Yes (3) Andersen and Anderson (1989): SV ��� therapists with 1 year of experience or less reported disclosing significantly less than therapists with 2���5 years and 10+ years experience Barrett and Berman (2001): EX ��� when therapists were instructed to increase their use of TSD, experienced therapist increased their TSD significantly more than less experienced therapists RESULTS MIXED but, if a relationship exists, it may be non-linear, with therapists with only the least (and, possibly, for male therapists, the most) amount of clinical experience disclosing less Robitschek and McCarthy (1991): SV ��� for male therapists, more experience correlated with less TSD No (3) Berg-Cross (1984): SV Simi and Mahalik (1997): EX Simone et al. (1998): SV Note. The numbers in parentheses indicate how many studies contributed to that question or possible answer. SV=survey EX=experiment. 64 J.R. Henretty, H.M. Levitt / Clinical Psychology Review 30 (2010) 63���77

other therapists, and one survey (Simi & Mahalik, 1997) found differences between psychoanalytic/psychodynamic and therapists of other orientations use of different types of, and reasons for, therapist self-disclosures. Although results were mixed, if there is a difference among therapists from various theoretical orientations, it appears that therapists from psychoanalytic and psychodynamic theories may disclose less, but that the difference may not be as large as theorized. 1.1.2. Independent variables related to the client and/or therapeutic process The studies reviewed in this paper examined the following independent client-related variables as they relate to therapist self- disclosure: age, diagnosis/type of disorder, symptomatology, and expectation of disclosure. Overall, results suggested that neither clients' age (Myers & Hayes, 2006 Simone et al., 1998), nor clients' expectation of therapist self-disclosure (Derlega, Lovell, & Chaikin, 1976 Peca-Baker & Friedlander, 1987 VandeCreek & Angstadt, 1985), had an interaction effect with therapist self-disclosure on clients' perceptions of, or responses to, the therapist. However, client diagnosis/type of disorder and therapist self-disclosure were shown to have a relationship, with therapists disclosing least to clients diagnosed with personality disorders (Mathews, 1989) or having weak ego-strength (Simone et al., 1998). Only one study examined clients' symptomatology in relation to therapist self-disclosure (Kelly & Rodriguez, 2007). It found that therapists self-disclosed more to clients with lower pre-therapy symptomatology. The only therapy- related independent variable investigated was alliance. The one study that examined alliance as an independent variable found that analogue clients with a strong therapeutic alliance perceived a self- disclosing therapist more favorably than clients with a weak alliance and a disclosing therapist (Myers & Hayes, 2006). 1.1.3. Independent variables related to issues of diversity Although few studies examined issues of diversity as they relate to therapist self-disclosure, this distinction seemed important, especially given thatfeminist and multicultural therapists tendto advocate theuse of therapist self-disclosure to reduce the power imbalance between therapists and clients, to decrease clients' feelings of shame, and to transmit feminist values from therapists to clients (Brown, 1994 Mahalik et al., 2000). Therefore, this subsection examines issues of sexual orientation, ethnicity, and gender. Only two studies examined the topic of sexual orientation as it pertains to therapist self-disclosure. One survey found that therapists of a feminist theoretical orientation were more likely to disclose their sexual orientation to clients than therapists of all other theoretical orientations (Berg-Cross, 1984). An analogue experiment by Atkinson, Brady, and Casas (1981) found that gay men rated male therapists who self-disclosed a gay sexual orientation more favorably than male therapists who did not disclose or disclosed a heterosexual orienta- tion, but that there was no difference between ratings of male therapists who did not disclose and those that disclosed heterosexual orientations. Due to diversity in research methodology, results across the two studies cannot be aggregated however, it may be that self- disclosing only makes a difference���that difference being positive��� when the therapist and client are both of a minority sexual orientation (Atkinson et al., 1981). Two studies (Edwards & Murdock, 1994 Myers & Hayes, 2006) coded for ethnicity (see Table 2) and found no differences among ethnic groups on clients' perceptions and responses to therapist self- disclosure. However, these studies had a relatively homogenous, mostly Caucasian sample. Six other studies specifically investigated ethnicity as it related to therapist self-disclosure with more varied samples, with three examining the effect of Mexican or Mexican- American culture, two examining the effect of African-American/Black culture, and one examining the effect of Asian-American culture. Although the number of studies is too few to be able to understand conclusively the particular effects each culture had on the impact of therapist self-disclosure, the results of the studies suggest that culture may interact with therapist self-disclosure. Studies that examined therapist self-disclosure and client gender, therapist gender, and/or gender pairing in the therapeutic dyad were compiled (see Table 3). Of all the diversity variables reviewed, it appears that only gender���of the client, therapist, and dyad���produced, in general, reliable results. Specifically, neither client gender, nor therapist gender, nor gender pairing, affected how much a therapist self-disclosed to a client. Additionally, gender did not reliably have an interaction effect with therapist self-disclosure on clients' perceptions of, and responses to, disclosing and nondisclosing therapists. 1.2. Dependent variables in the therapist self-disclosure research This section is subdivided into three parts. The first subsection discusses articles that examined dependent variables associated with therapists. The therapist-related dependent variables in the literature tend to be ratings of the therapist's demeanor and performance in therapy. The second subsection investigates dependent variables related to clients (such as clients' feelings toward the therapist) or the process of therapy (such as alliance, as it is affected by disclosure). The third subsection reviews studies that investigated variables related to client self-disclosure. The division of this section (dependent variables) into these three subsections (therapist-related dependent variables, client-related dependent variables, and dependent variables related to client self-disclosure) is indicative of the distinction among clients' perceptions���that is, client ratings of which the therapist is the subject (e.g., ���my therapist is attractive���) clients' feelings���that is, client ratings of which the client is the subject (e.g., ���I am attracted to my therapist���) and client disclosure measures���ratings which, except in the case of client willingness to disclose, are done by someone other Table 2 Is there a relationship between client ethnicity and therapist self-disclosure (TSD)? (8). Possible answers Studies Summary Yes (4) Cashwell, Shcherbakova, and Cashwell (2003): ASV ��� clients indicated preference for TSD when therapist was of different ethnicity African-American clients indicated higher preference for TSD than Caucasian clients about personal feelings, sexual and professional issues, and success/failure Wetzel and Wright-Buckley (1988): AEX ��� Black clients paired with Black therapists showed preference to high TSD therapy condition Black clients paired with White therapists showed preference to low TSD therapy condition RESULTS MIXED, but, if a relationship exists, clients of Mexican cultures may prefer nondisclosure, whereas African-American/ Black clients may prefer TSD Cherbosque (1987a): ASV ��� American clients expected more TSD than Mexican clients Cherbosque (1987b): AEX ��� Mexicans showed preference for No TSD therapy Americans did not No (4) Borrego, Chavez, and Titley (1982): AQEX ��� Mexican-American vs. Anglo-American Kim et al. (2003): AQEX ��� Asian values Edwards and Murdock (1994): SV ��� Caucasian vs. non-Caucasian Myers and Hayes (2006): AEX ��� general ethnicity Note. The numbers in parentheses indicate how many studies contributed to that question or possible answer. SV=survey ASV=analogue survey AEX=analogue experiment AQEX = analogue quasi-experiment. 65 J.R. Henretty, H.M. Levitt / Clinical Psychology Review 30 (2010) 63���77