xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001916891
007 cr mnu|||uuuuu
008 071114s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001799
Handelsman, Jessica B.
Linking pretreatment therapist characteristics to the therapeutic alliance in youth treatment :
b an examination of professional burnout, counseling self-efficacy and gender role orientation
h [electronic resource] /
by Jessica B. Handelsman.
[Tampa, Fla] :
University of South Florida,
ABSTRACT: The present study investigated three pretreatment therapist characteristics (professional burnout, counseling self-efficacy, and gender role orientation) in relation to the therapeutic alliance within the context of youth treatment. It was hypothesized that the emotional exhaustion and depersonalization dimensions of burnout would be negatively associated with the alliance, while the personal accomplishment dimension of burnout and counseling self-efficacy would be positively associated with the alliance. In addition, it was hypothesized that androgynous therapists would have superior alliances, relative to stereotypically masculine or feminine therapists. Participants were 42 pairs of therapists and youth clients. Prior to intake, therapists completed the Maslach Burnout Inventory -- Human Services Survey (MBI), a modified version of the Counselor Self-Efficacy Scale (CSES-M), and the Bem Sex-Role Inventory (BSRI). Clients and therapists completed parallel versions of the Child Therapy Bond Scale (CTBS) following the third session. As hypothesized, results indicated that depersonalization and personal accomplishment were significantly related, in the expected directions, to therapist ratings of the alliance. Other hypotheses were not supported. Future research directions and potential implications of these findings for professional training, service delivery, and quality management in mental health organizations are discussed.
Thesis (M.A.)--University of South Florida, 2006.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 91 pages.
Adviser: Marc Karver, Ph.D.
Gender role orientation.
t USF Electronic Theses and Dissertations.
Linking Pretreatment Therapist Characteristics to the Therapeutic Alliance in Youth Treatment: An Examination of Professional Burnout, Counseling Self-Efficacy and Gender Role Orientation by Jessica B. Handelsman A thesis submitted in partial fulfillment of the requirements for the degree of Master of Arts Department of Psychology College of Arts and Sciences University of South Florida Major Professor: Marc Karver, Ph.D. Vicky Phares, Ph.D. Joseph Vandello, Ph.D. Date of approval: October 4, 2006 Keywords: therapeutic alliance, therapist characteristics, professional burnout, counseling self-efficacy, gender role orientation Copyright 2006, Jessica B. Handelsman
Table of Contents List of Tables iii List of Figures iv Abstract v Introduction 1 Overview 1 Therapeutic Alliance 2 Professional Burnout 5 Counseling Self-Efficacy 10 Gender Role Orientation 12 Masculinity 14 Femininity 16 Androgyny 17 Current Study 22 Methods 23 Participants 23 Measures 29 Therapist Background Form 29 Case Information Form 30 Maslach Burnout Inventory 30 i
Counselor Self-Efficacy Scale-Modified Version 32 Bem Sex Role Inventory 33 Child Therapy Bond Scale 34 Procedures 35 Results 37 Descriptive Statistics 37 MBI-Emotional Exhaustion 37 MBI-Depersonalization 37 MBI-Personal Accomplishment 38 Counselor Self-Efficacy Scale-Modified Version 38 Bem Sex Role Inventory 38 Child Therapy Bond Scale 40 Hypothesis Testing 43 Post Hoc Analyses 48 Discussion 50 References 67 Appendices 80 ii
List of Tables Table 1 Therapist Demographic and Background Information 25 Table 2 Client Demographic and Background Information 28 Table 3 MBI Subscale Score Classifications 31 Table 4 MBI Subscale Correlation Matrix 31 Table 5 Descriptive Statistics for Independent Variable Measures 41 Table 6 Gender Role Orientation Rates 42 Table 7 Descriptive Statistics for the CTBS 43 Table 8 Intercorrelations between Independent and Dependent Variables 44 Table 9 Results for Each Gender Role Orientation Category 45 iii
List of Figures Figure 1 Theoretical Model 23 iv
Linking pretreatment therapist characteristics to the therapeutic alliance in youth treatment: An examination of professional burnout, counseling self-efficacy and gender role orientation Jessica B. Handelsman ABSTRACT The present study investigated three pretreatment therapist characteristics (professional burnout, counseling self-efficacy, and gender role orientation) in relation to the therapeutic alliance within the context of youth treatment. It was hypothesized that the emotional exhaustion and depersonalization dimensions of burnout would be negatively associated with the alliance, while the personal accomplishment dimension of burnout and counseling self-efficacy would be positively associated with the alliance. In addition, it was hypothesized that androgynous therapists would have superior alliances, relative to stereotypically masculine or feminine therapists. Participants were 42 pairs of therapists and youth clients. Prior to intake, therapists completed the Maslach Burnout Inventory Human Services Survey (MBI), a modified version of the Counselor Self-Efficacy Scale (CSES-M), and the Bem Sex-Role Inventory (BSRI). Clients and therapists completed parallel versions of the Child Therapy Bond Scale (CTBS) following the third session. As hypothesized, results indicated that depersonalization and personal accomplishment were significantly related, in the expected directions, to therapist ratings of the alliance. Other hypotheses were not supported. Future research directions and potential implications of these findings for professional training, service delivery, and quality management in mental health organizations are discussed. v
1 Introduction Overview The purpose of the present study is to examine the relationships between several therapist variables and the therapeutic alliance within the context of child and adolescent mental health treatment. Empirical research has identified the therapeutic alliance to be among the most robust predictors of proxima l and distal treatment outcomes for both youth and adult clients (Karver, Handelsma n, & Fields, 2006; Lambert & Barley, 2002; Safran & Muran, 2000; Shirk & Ka rver, 2003). At this point, however, there is limited research to inform the fiel d about specific vari ables that impact the development of therapeutic alliances w ith youth clients. Research on common process factors indicates that certain therapist traits and behaviors are likely to influe nce the quality of relationshi ps with clients (e.g., Creed & Kendall, 2005; Karver et al., 2006). For inst ance, studies in the adult treatment field indicate that therapists who fo rm strong alliances with client s tend to present as flexible, honest, respectful, trustworthy, confident, warm, interested, and open (Ackerman & Hilsenroth, 2003). Furthermore, specific th erapist personality traits such as neuroticism, rule consciousness, independence, dominance, social control, perfectionism, and impression management have demonstrated negative relationships with the alliance in child and adolescent treatment (Doucette Boley, Rauktis, & Pleczkowski, 2004). Yet, which variables facilitate ther apists abilities to demonstrate alliance-enhancing versus
2 alliance-hindering traits and behaviors remains unclear. The current study examines three pretreatment therapist variables pr ofessional burnout, counseling self-efficacy, and gender role orientation that may be im plicated in the formation of therapeutic alliances with youths. Before further discussing these factors, background research on the therapeutic alliance in youth mental health treatment will be reviewed. Therapeutic Alliance Consideration of the therapeu tic alliance first appeared in the adult psychotherapy literature. In his early theoretic work, Fr eud discussed the importance of developing a collaborative relationship between the anal yst and the patient (Meissner, 1996). He focused primarily on the transferential quali ty of the relationship, which he saw as essential for therapeutic change (Meissner, 1996; Safr an & Muran, 2000). Several alternative conceptualizations of the allian ce emerged, as researchers strived to better account for the common mechanisms of change across treatment approaches (Safran & Muran, 2000). One reformulation of the alliance, put fo rth by Bordin (1979), has earned a great deal of attention in the a dult field and served as the foundation for many subsequent attempts to quantify the therapeutic relations hip. Bordin proposed that the alliance is both a facilitator of treatment and a change mechanism in itself (Shirk & Karver, in press). His model, consisting of thre e interrelated components, emphasizes the complicated, dynamic, and multidimensional nature of the alliance. The first component, Tasks, represents the specific activities that therapists a nd clients engage in throughout treatment. Bordin highlighted the importance of joint collaboration on these activities.
3 The second component, Goals, represents the basic objectives targeted by a given treatment. In Bordins view, mutual agreement or consensus on goals is vital to the treatment process. Lastly, Bond represents the affective component of the therapeutic relationship, which allows clients to f eel understood, respected, and valued by their therapists. Bordin suggested that the quality of the em otional connection between a therapist and client mediates their collective ability to negotiate tasks and agree upon goals. While not all researchers have adopted Bordins framework, most agree that the alliance is comprised of both relational and technical asp ects (Meissner, 1996; Safran & Muran, 2000). More specifically, several assumptions about the therapeutic alliance are relatively universal across theo retical models. First, the al liance is thought to play a functional and important role in the treat ment process (Lambert, 2004; Meissner, 1996; Safran & Muran, 2000). Second, it is assumed that the alliance begins to form upon initial referral to treatmen t. By extension, developing the groundwork for a strong alliance in the early stages of treatment is regard ed as beneficial (Lambert, 2004). Third, the alliance is considered to be reciprocal a nd mutual that is, both clients and therapists bring to treatment individual characteristics that influence the development of the alliance (Lambert, 2004; Safran & Muran, 2000). Finally, the alliance is pres umed to be dynamic and malleable, as it develops throughout the co urse of treatment and may be shaped by specific therapist and client behaviors (L ambert, 2004; Meissner, 1996; Safran & Muran, 2000).
4 While the adult mental health field ha s debated and studied the therapeutic alliance for many years with over 2000 arti cles published since 1977 (Horvath & Bedi, 2002) the focus on this construc t is relatively new in the child and adolescent literature. There has been a recent increase in the number of studies examining the alliance in clinical samples of youths, as more researchers have come to recognize that the relevance of therapeutic relationships is not limited to adult treatment. In fact, it has been suggested that the alliance ma y be particularly important in working with youths, as child and adolescent clients typical ly are not self-referred and often enter into treatment unaware of their problems, in conflict with their primary careg ivers, and/or resistant to change (DiGiuseppe, Linscott, & Jilton, 1996; Shirk & Karver, 2003). Developing a strong therapeutic relationship with young clients may lessen resistance to treatment and facilitate engagement by providing a stab le, accepting and suppor tive context within which therapy may take place. This theory was upheld in a recent meta-analytic review of 23 studies, wherein Shirk and Karver (2003) showed that the therapeutic relationship is related to distal treatment outcomes for child ren and adolescents. With effect sizes ranging from .21 to .26 (which are comparable to those reported in th e adult literature), the therapeutic relationship represents one of the strongest predictors of treatment outcomes for children and adolescents. Given that the alliance appears to serve a vital function in the treatment of youths, it would be helpful to know which factor s contribute to its development. As aforementioned, therapist interpersonal quali ties (e.g., warmth) and personality traits (e.g., neuroticism) have been shown to predict the quality of alliances formed in mental
5 health treatment. Yet, few studies have exam ined variables that may facilitate or hinder therapists abilities to demonstrate these alliance-enhancing characteristics. The following sections discuss a theoretical model of the process by which therapists levels of professional burnout, counseling self-efficacy, masculinity, and femininity may be implicated in the formation of therapeutic alliances with children and adolescents. To date, little research has inve stigated how these factors are related to one another and no studies have examined their links to th e therapeutic alliance in youth treatment. Understanding the relationships between these variables may provide a better understanding of treatment processes, wh ile also laying the groundwork for training clinicians and improving mental he alth interventions for youths. Professional Burnout As with many occupations, working in the mental health field can be both rewarding and demanding of professionals emoti onal, cognitive, and physical resources. Within the context of treatment, therapists personal resources are not only directed toward identifying and accommodating their clie nts individual needs, but also towards self-monitoring their own thoughts, feelings, and behaviors in clinical situations, particularly those situations that elicit cognitive dissona nce, emotionality, or other countertransference reactions. In child and adolescent treatment, therapists personal resources are frequently also devoted to developing positive working relationships with their clients parents, teachers, and other caregivers or service providers, who may be relied upon for the purposes of providing info rmation, scheduling sessions, transporting clients to and from sessions, facilitating in terventions during and between sessions, and
6 monitoring clients safety and compliance with treatment recommendations (Fields, Handelsman, Karver, and Bickman, manuscr ipt in preparation). However, the professional demands on service providers ex tend beyond their therap eutic roles. Today, practitioners often struggle to reconcile the conflicting interests of clients, referral sources (e.g., parents, teachers, social service agen cies), program administrators, insurance companies, and other vested parties (Rupert & Morgan, 2005). The shifting economy and the rise of managed healthcare have put greater financial pressure on service providers to increase their caseloads and shor ten the length of treatm ent, while generating rapid and long-lasting clinical results (Rupert & Baird, 2004; Rupert & Morgan, 2005). In addition, changes in professional and legal guidelines regarding assessment, documentation, and reporting, coupled with downsizing within organizations due to financial constrictions, have increased the demands placed on mental health service providers (Rupert & Baird, 2004; Rupert & Morgan, 2005). Given the growing pressures therapists face, it is not surprising that the phenomenon of professional burnout a uni que response syndrome arising out of chronically heightened job demands (Zoha r, 1997, p.101) has received increased attention within the human services field ove r the past twenty y ears (Rupert & Morgan, 2005). Professional burnout was estimated to af fect as many as one-third of practicing psychologists in the 1980s (Ackerley, Burne ll, Holder, & Kurdek, 1988). While more contemporary prevalence rates have not been published, a number of recent studies indicate that burnout continue s to be a significant issu e for psychologists and other
7 service providers within mental health sett ings (Bakker et al., 2006; Rosenberg & Pace, 2006; Rupert & Baird, 2004; Rupert & Morgan, 2005). The literature on this topic has been hi ghly influenced by the theoretical and empirical work of Maslach and her colleagues. In their original model, Maslach and Jackson (1986) conceptualized professional burnout on a tri-dimensional continuum. The first dimension, Emotional Exhaustion (EE), refers to a depletion of emotional and psychological resources available to perform in ones professional role, resulting in fatigue and/or distress (Schaufeli & Enzman, 1998). The second dimension, Depersonalization (DP), represents the development of a cognitive bias towards making negative, impersonal, and de humanizing attributions about the recipients of ones services (Schaufeli & Enzman, 1998). The third dimension, Personal Accomplishment (PA), refers to positive self-evaluations regarding ones ability to perform his/her professional roles competently and with ease. PA also refers to feelings of fulfillment and satisfaction regarding ones work or im pact on clients (Schaufeli & Enzman, 1998). According to this model, burnout is viewed not as a collection of individual symptoms, but as a developmental process that involves an interaction between internal and external factors, and thus fluctuates over the course of ones career (Bakker, Van Der Zee, Lewig, & Dollard, 2006; Corey & Corey, 1998; Evan s et al., 2006; Kestnbaum, 1984; Rosenberg & Pace, 2006). Extensive research much of which has utilized the Maslach Burnout Inventory (MBI; Maslach & Jackson, 1981, 1996), a measure comprised of three subscales, each of which represents one of the three burnout dimensions has demonstrated support for this model (Mas lach, Jackson, & Leiter, 1996).
8 Levels of burnout appear to vary widely both within and across samples, based on age (Vredenburgh, Carlozzi, & Stein, 1999), le vels of training/experience (Farber, 1990; Cushway & Tyler, 1996), and professional roles (Boice & Myers, 1987; Dupree & Day, 1995; Farber, 1990; Finnoy, 2000; Onyett, Pi llinger, & Muijen, 1997; Radeke & Mahoney, 2000; Raquepaw & Miller, 1989; Vr edenburgh, Carlozzi, & Stein, 1999). One consistent finding across studi es has been the relationship between burnout and work setting, with therapists in the private sector reporting less burnout than those who are agency-employed (Ackerley et al., 1988; Farber, 1983; Hellman & Morrison, 1987; Raquepaw & Miller, 1989; Rupert & Morg an, 2005; Vredenburgh, Carlozzi, & Stein, 1999). Ackerley, et al. (1988) reported that the modal burned-out clinician in their sample was young, had a low income, engaged in little individual therapy, experienced feelings of lack of control in the therapeu tic setting, and felt over-c ommitted to clients. In two more recent surveys (conducted in 1999 and 2001) of licensed psychologists whose primary employment was in private prac tice, Rupert and Baird (2004) found that high involvement with managed care was asso ciated with working longer hours, having more client contact, receiving less supervision, reporting more negative client behaviors, experiencing more stress, being less satisfied with their incomes, and having higher levels of emotional exhaustion. Research has demonstrated that burnout can have extensive physical, emotional, interpersonal, and attitudi nal implications for profe ssionals, including poor physical health, depression, problematic interpersonal relations, negative at titudes regarding job satisfaction, unproductive work behaviors, a nd job turnover (Kahill, 1988), as well as
9 progressive loss of idealism, energy, purpose, and concern due to work-related strain (Edelwich & Brodsky, 1980, p. 14). As such, it is not surprising that levels of burnout were predictive of reported intentions to leave psychotherapy for individuals in Raquepaw and Millers (1989) sample of practicing psychologists. The potential implications of burnout extend beyond the pers onal costs for indivi dual professionals. Burnout also can have a negative impact on organizations, as they confront problems associated with staff turnover and shortages, as well as excessive caseloads for remaining staff (Evans et al., 2006). Moreover, client s may be affected by therapist burnout, as the quality of services they receive may suffer (Rupert & Morgan, 2005). While no research has looked directly at the relationship between professional burnout and the therapeutic alliance, it reasons that therapists w ho become over-extended in trying to meet the many demands associat ed with their professional roles may have inadequate resources availabl e for facilitating treatment a nd fostering the therapeutic relationship. More specifically, emotional exhaustion may undermine therapists abilities to convey warmth, trustworthiness, con cern, engagement, and other interpersonal characteristics shown to promote collaborat ion, consensus, and a therapeutic bond with clients (Ackerman & Hilsenroth, 2003). In addition, emotional distress may interfere with therapists abilities to se lf-monitor and attend to client s behavior du ring sessions. Manifestations of depersonaliz ation such as the devel opment of negative, callous, cynical, or ambivalent attitudes towards clients could lead therapists to demonstrate poor motivation, decreased investment, and negativ e emotionality with re spect to clients. Therapists who lack positive professional attitu des and prosocial approaches to treatment
10 may be less able to elicit engagement and pa rticipation from clients. Therapists who experience a diminished sense of personal accomplishment may also have difficulty forming strong therapeutic alliances. Therapis ts negative self-perceptions and attitudes regarding their clinical competence, therap eutic abilities, and act ual performance may lead to increased anxiety, frustration, pe ssimism, or hopelessness th at is apparent to clients. It reasons that all three dimensions of professional burnout are likely to influence therapists behavior in ways that could je opardize the therapeutic relationship. It is hypothesized that higher levels of emotional exhaustion and depersonalization, as well as lower levels of personal accomplishment, will be associated with more negative ratings of the therapeutic alliance. Counseling Self-Efficacy Related to professional burnout is the con cept of self-efficacy -conceptualized as ones perceived capacity to perform a particular action (Larson & Daniels, 1998). Research has shown that this factor is imp licated in peoples acti ons, decisions, effort expenditure, perseverance, thought patterns, and levels of stress (Bandura, 1986, 1989). Counseling self-efficacy (CSE) has been de fined as ones beliefs or judgments about her or his capabilities to effectively counsel a client in the near future (Larson & Daniels, 1998, p.180). Therapists levels of CSE may be influenced by a number of different factors, including self-perceptions regarding th eir knowledge of psychological principles; their abil ities to monitor and control th eir own thoughts, feelings and behaviors in order to adapt to clients n eeds; their familiarity with assessment and treatment strategies, and their abilities to employ these techni ques effectively; as well as
11 their competence to perform in a manner that is congruent with ethi cal and professional standards. Counseling self-efficacy has been shown to have significant implications for mental health treatment. Specifically, empiri cal research has demonstrated that CSE has a negative relationship with levels of therap ist anxiety (Friedlander et al., 1986; Larson & Daniels, 1998; Larson, Suzuki, Gillespie, Po tenza, Bechtel, & Toulouse, 1992) and a positive relationship with actual counseling skills and performance, based on therapists self-ratings (Larson et al., 1992, Larson & Daniels, 1998; Wester, Vogel & Archer, 2004), supervisor ratings (Larson & Daniels, 1998), and independent ratings (Larson et al., 1992; Munson, Stadulis & Munson, 1986). Larson and Daniels (1998) suggest that low CSE may lead to avoidance, unwilli ngness to take risks, and diminished perseverance. In addition, Larson, Cardwell, and Majors (1996, unpublished) reported that CSE had a modest, but significant, pos itive correlation with therapist burnout (as cited in Larson & Daniels, 1998). Research has not yet examined CSE in relation to the therapeutic alliance, but it reasons that therapists with high CSE may be more able to adapt to clients individual needs, to convey self-confidence, and to othe rwise exhibit the interpersonal qualities that foster positive therapeutic relationships with clie nts. On the flip-side, therapists with low CSE may demonstrate occupational stress, poor confidence, lack of expertise, diminished motivation, and other behaviors that could interfere with their abilities to respond effectively to clients in order to form st rong relationships. As such, it is hypothesized that levels of CSE will be positively co rrelated with the therapeutic alliance.
12 Interestingly, evidence suggests that men generally have lower self-efficacy for traditionally female occupations (Bonett, 1994), while women generally have lower selfefficacy for traditionally male occupations (Bonett, 1994; Matsui, 1994). Yet, several studies have reported that levels of CSE in clinical trainees and prof essional therapists do not differ by sex (Larson et al., 1992; Potenza, 1990 as cited in Larson, 1998). Perhaps more important than the actual sex-ratio within a pr ofession is how congruent it is with ones gender role orientation. Gender Role Orientation Gender-roles can be defined as the totality of social and cultural expectations for boys/girls, men/women in a particular society at a particular time in history (Byer, Shainberg, & Galliano, 1999, p.345). Traditional gender-norms dictate that males value and strive for personal achievement, power and status, goal-attainment, self-reliance, competition, and restriction of emotionality (Freudenberger, 1990; Heppner & Gonzales, 1987; Wester & Vogel, 2002), while fema les value and strive for closeness, supportiveness, caring, interpersonal warmth and understanding (Romans, 1996). Yet, gender socialization has evolve d over time, reflecting a grad ual cultural shift towards more egalitarian sex-roles. Manifestations of this change are seen in the greater numbers of women adopting professional roles (Jome & Tokar, 1998), including in the mental health service field (American Psychologi cal Association Research Office, 2003). Historically, the majority of psychotherapists were male (American Psychological Association Research Office, 2003), while the majority of clients were female (Heppner
13 & Gonzales, 1987; Kohout & Wicherski, 1999; Pleck, 1987). For over three decades, however, women have represented the majority of students in undergraduate, masters, and doctoral level psychology, counseling, a nd social work programs. Although men remain in the majority of administrative, academic, and research positions, women have increasingly outnumbered men in mental hea lth service positions since the late 1970s (American Psychological Association Res earch Office, 2003; Kadushin, 1976). In addition, it has become more normative for males to be mental health consumers (Pleck, 1987). These shifts may have important im plications for the field of psychology. Changes in gender socialization have para lleled changes in conc eptualizations of, and approaches to, mental health treatment. In the era of Freudian psychoanalysis, exchanges between the client and psychothera pist were often portrayed as cold and formal (Freudenberger, 1990; Meissner, 1996). Approaches to treatment were based on a framework that reinforce[d] traditional male role attributes through the expectations that therapists only reflect patients feelings and give no evidence of their own personal feelings, appearing strong and being sile nt (Freudenberger, 1990, p.340). While more contemporary theoretical orientations such as cognitive-behavioral and humanistic vary in their definitions of the therapist s role, there seems to be greater emphasis on eliciting and maintaining clie nts engagement in treatme nt by creating an accepting, warm, and trusting atmosphere (Lambert, 2004) Consequently, therapy today may be viewed as a stereotypically feminine activity in the sense that it is often associated with supportiveness, emotional responsiveness, and in terpersonal sensitivity, rather than more
14 traditionally masculine ideals, such as self-re liance, restriction of emotional expression, status, and competition for power (Harvey & Hansen, 1999; Wester & Vogel, 2002). On the other hand, there are aspects of providing therapy that remain more aligned with stereotypically masculine qualities such as demonstrating authorita tiveness, expertise, goal-directedness, and assertiv eness (Harvey & Hansen, 1999). All things considered, it seems that therapy is not easily classified as a gender-typed occupa tion. Thus, regardless of sex, therapists with highly masculine or highly feminine orientations may have difficulty when called upon to perform in ways that are incongruent with their gender role traits. Masculinity Several theories have been put forth regarding the therapeutic implications of having a stereotypically masculine gender role orientation as a therapist. For instance, Wester and Vogel (2002) suggest that masculine gender socialization, emphasizing success and competition, may drive some therap ists to assert their clinical prowess, rather than focusing on the clients issues, a ssume authority and/or try to assert control within interpersonal relations hips (372). Related, Heppner and Gonzales (1987) suggest that therapists with a masculine gender role orientation may be compelled to assert their status and create a power differe ntial with clients, not for therapeutic reasons, but simply for the sake of control (35) It follows that stereotypically masculine therapists may have more difficulty forming positive alliances w ith child and adolescent clients, as their behavior may be perceived as domineering, co ld, threatening, and/or patronizing. As
15 aforementioned, Doucette, et al. (2004) found that therapists with higher levels of dominance, independence, and social contro l had more negative alliances with youth clients. If these traits are interpreted as re presentative of stereotypically masculine traits, these findings provide additional support for th e theory that therapists with masculine gender role orientations may have less success in forming strong alliances with children and adolescents. It has also been proposed that stereot ypically masculine therapists may be less inclined to express empathy, warmth, and in timacy, particularly towards male clients (Heppner & Gonzales, 1987; Sher, 2001), as feelings of concern and expressions of affect may be perceived as incongruent with masculine norms. According to Heppner and Gonzales (1987), [If] the counselor is uncomfortable accepting and expressing his own emotions, he may inhibit, consciously or unconsciously, the clients expression of emotion (34). In support of this view, Hayes (1984) found that male psychology trainees scoring higher on the Restricted Emotionality and Restrictive Affectionate Behavior Between Men subscales of the Gender Role Conflict Scale reported less empathy for and more interpersonal difficulties with both gay and highly emotional male clients. These findings were replicated by Wisch and Mahalik (1999). Therapists who have difficulty with the emotional aspects of treatment or experi ence discomfort with clients who demonstrate untraditional gender role traits, may appear uncommitted, unsympathetic, or insensitive to their clients needs. As a result, stereotypically masculine therapists may be less able to foster the therapeutic alli ance with children and
16 adolescents in treatment. Empirical research is needed in order to further evaluate this theory. Femininity Despite the increasing number of wo men in the mental health field, a comprehensive review of the mental health service literature produced no empirical studies that specifically examine female therap ists gender role orientations. However, research outside the treatment literature has shown femininity to be associated with expressive behavior and a humanistic or ientation (Harvey & Hansen, 1999), both of which may facilitate development of the th erapeutic alliance with youth clients. For instance, MacGeorge, Clark, and Gilliha n, (2002) found that female communication students produced emotional support messages with a higher level of person-centeredness and reported greater self-efficacy in the domain of providing emotional support, compared to their male counterparts. The authors suggest that [ h]ighly person-centered messages reflect a more complex set of percep tions and intentions, pursue broader sets of interaction goals, and in an important sens e, do more work than less person-centered messages (18). It is notable that self-e fficacy mediated approximately 30% of the sexrelated variance in person-centeredness. If these findings generali ze to mental health professionals, female therapists (or therapists with high levels of femininity) may also communicate more person-centered emotional support messages and have greater selfefficacy in this area than male therapists (or therapists with low levels of femininity). Moreover, therapists perceptions of their ow n abilities may play an important role in how they actually behave.
17 While having feminine traits may be an asset in providing treatment, therapists with only these traits to draw upon may be li mited when faced with clinical situations that call for more stereotypically masculine responses. According to Abramowitz and Abramowitz (1976), decision-making, risk-taki ng, and other aspects of the therapeutic role that are more stereotypically masculin e may activate gender-related anxieties for sex-typed female therapists. Related, Carls on (1987) suggests that, in working with male clients, female therapists have an obligat ion to grow beyond the sex role that traps women into limiting behavior that in turn does not challenge the male client and presumes a power imbalance in his favor and a caregiver role for her (47). Given that gender socialization has greatly evolved since these views we re put forth several decades ago, it is pertinent to establish how femininity is imp licated in contemporary mental health treatment. As with stereotypically masculine therapists, it may be that stereotypically feminine therapis ts, regardless of sex, are rest ricted in thei r abilities to adapt to the diverse need of clients in order to form strong alliances. Androgyny Given that both instrumental (masculine) and expressive (feminine) therapist traits are likely to facilitate treatment with different clients, it may be most accurate to view therapy as an androgynous activity. Research on androgyny suggests that masculinity and femininity are independent di mensions, rather than opposite ends of the same continuum (Scher, Stevens, Good, & Eich enfield, 1987), that may be integrated and balanced within a single person (Kravetz & Jones, 1981). Research has shown that androgynous individuals are less likely than sextyped individuals to pr efer activities that
18 are congruent with traditiona l gender roles, and tend to re port less discomfort when required to perform sex-inappropriate beha viors (Kravetz & Jones, 1981). Assuming that therapy requires gender role flexibil ity, it reasons that androgynous individuals may be best equipped for this profession. Harvey and Hansen (1999) suggest th at the combination of masculine and feminine traits allows androgynous individuals to select from a broader repertoire of either type traits for the skills necessary at the time (106). In support of this theory, Kravetz and Jones (1981) found that androgynous individuals we re more able than sextyped individuals were to adapt their behavior in response to varyi ng situational demands. Similarly, Nevill (1977) reported that the availa bility of multiple roles is related to greater skill in social behavior. It follows that androgynous therapists may be better able to shift their therapeutic style in orde r to accommodate clients i ndividual differences. Related, Fong and Borders (1985) found, based on independe nt ratings of training counselors, that gender role orientation had a significant effect on counsel ing skills scores and response effectiveness before and after skills traini ng, while therapists sex was not significantly related to counseling performance. Andr ogynous trainees were significantly more effective prior to training, although this group difference did not remain significant following training. These findings support the notion that levels of masculinity and femininity may be more important than therap ists sex, and that clin ical training may be able to modify less effective gender-r elated traits and behaviors. Research also has demonstrated that androgyny is positively correlated with multiple indices of adaptive psychosocial func tioning, including: self-esteem, behavioral
19 flexibility, and interpersonal adjustment (Harvey & Hansen, 1999), as well as selfactualization, spontaneity, self -regard, self-acceptance, fee ling reactivity, and capacity for intimate contact (Nevill, 1977). It reasons that, compared to masculine and feminine therapists, androgynous therapists may be bett er adjusted, have gr eater confidence in their abilities to meet the diverse needs of clients (higher CSE), a nd experience less stress and/or conflict (professional burnout) with respect to managing the various demands of their professional roles. Little research has addressed the potential implications of therapists gender role orientations and no studies examining this vari able in relation to the therapeutic alliance were found. In fact, very little theoretical work has been published in this area. Yet, this topic warrants attention for multiple reasons. First, gender role orientation may be relevant to who becomes a therapist. It is possible that a self-selection process occurs whereby individuals with less traditional gender-related traits tend to enter counseling positions. Al ternatively, education, training, supervision, and experience may guide therapis ts professional deve lopment towards less traditional gender role orientations. If either case is true, the current distribution of therapists gender roles should be skewed, reflecting a greater number of androgynous therapists. Empirical research suggests that male psychologists and trainees experience the same gender role socialization as ot her males (Heppner & Gonzales, 1987; Wester, Vogel, & Archer, 2004; Wisch & Mahalik, 1999). While female therapists have not been studied in this manner, it is likely that they too are expos ed to the same gender role
20 socialization as other females. However, whether individuals in counseling roles internalize and act out sex-type d gender-norms within the cont ext of therapy is unclear. In a study of male psychologists, Ha rvey and Hansen (1999) found that the majority of their sample (54%) reporte d an androgynous gender role within the professional setting, while 6% reported an undi fferentiated gender role (low femininity and low masculinity), 25% reported a femi nine gender role, and only 15% reported a masculine gender role. Over half of the therapists who described themselves as androgynous in their professional roles, also described themselves as androgynous in their personal roles. More research is need ed in order to determine if these results are representative of male therapists overall. Fu rthermore, research is needed to investigate the distribution of gender role orientations in female therapists. Second, gender role orientations may influence how therapists conceptualize and approach treatment, and therefore may im pact the interpersonal dynamics between therapists and their clients. Children and adolescents enter into therapy with a range of gender-related traits manifested in their at titudes, expectations, feelings, and behaviors that may or may not be directly linked to their presenting problems. Gender-related issues may be particularly salient in child and adolescent clients, as interpersonal attachment, identity formation, conflict with authority, sexuality, and social comparison, are all prominent issues during early devel opment (Steinberg, 1996). Thus, therapists may be called upon to exhibit eith er stereotypically feminine behaviors or stereotypically masculine behaviors. Therapists low in femini nity and/or masculinity may be less able to adapt their therapeutic styles in order to accommodate the individual needs of their youth
21 clients. As such, therapists gender role or ientations may either facilitate or undermine the formation of strong alliances during the treatment process. A third incentive for studying therapists gender role orientations is that education, training, and clinical supervision may intermittently reinforce and contradict gender-norms, but not address gender role issues directly (Wes ter & Vogel, 2002). Consequently, therapists may be susceptible to gender role conflict a psychological state that occurs when rigid, sexist, or restrictive gender roles, learned during socialization, result in pers onal restriction, devaluation, or violation of others or self (ONeil, 1990, p.25). In samples of male therapists, gender role conflict has been negatively associated with c ounseling self-efficacy (Wester, Vogel, & Archer, 2004) and positively associated with countertransference reactions that interfered with treatment (Hayes, 1984). Related to these findings Kadushin (1976) found that male social workers who experienced conflicts between gender identity and occupational status indicated that this gender role conflict affected their relations hips with colleagues, clients and the community in general. They repor tedly adjusted to these problems by choosing fields of practice and methods that are mo re stereotypically masculine, or by moving toward the administrative level of the professi onal hierarchy. No research on gender role conflict in female therapists was found. Without knowledge of how to incorporate ge nder into their prof essional identities, therapists may experience gender role c onflict, which in turn may contribute to professional burnout, diminish counseling self -efficacy, and interfere with therapists
22 abilities to establish strong relationships with their clients. Additional research may help guide supervisors in how to address gender-re lated issues with their clinical trainees. In conclusion, it is hypothesi zed that therapists with both masculine and feminine traits to draw from (i.e., androgynous therapis ts) will have stronger therapeutic alliances with their clients, compared to masculine and feminine therapists. It is also hypothesized that androgyny will be positively correlated with CSE and negatively correlated with professional burnout. Furthermore, CSE a nd burnout are expected to mediate the relationship between gender role orie ntation and the therapeutic alliance. Current Study: In the present study, therapists profe ssional burnout, counseling self-efficacy, and gender role orientations are examined in relation to one another and the therapeutic alliance. The study was guided by the aforem entioned model (see Figu re 1) representing the theoretical relationship between these factor s. (Note that Thera pist Behaviors and Proximal and Distal Treatment Outcomes are included in the model to provide greater context, but these variables we re not measured in the present study.) A summary of the hypotheses tested is provided in Figure 1.
23 Figure 1. Theoretical Model COUNSELING SELF-EFFICACY PROFESSIONAL BURNOUT THERAPEUTIC ALLIANCE THERAPIST BEHAVIORS PROXIMAL & DISTAL TREATMENT GENDER-ROLE ORIENTATION
24 Methods Participants Forty-two pairs of primary therapists and youth clients (ages 6-17), who began treatment together between July 2004 and Ap ril 2005, were included in the present study. This sample was drawn from a non-profit or ganization that provides mental health services in a Midwestern region of the continental United Stat es. The organization serves approximately 385 youths per day through a variety of community-based programs, including: Intensive Family Preservation (I FP), Therapeutic Foster Care (TFC), four residential treatment facilities/therapeutic group homes (RTF/TGH), and an outpatient eating disorder clinic (OPC). The vast major ity (88%) of therapistclient dyads included in the present study were involved with eith er IFP or TFC, both of which provide inhome mental health services to youths and their families. Demographic characteristics of the therap ist sample are shown in Table 1. Most therapists were under the age of 40, female, white/Caucasian, and parents, respectively. Furthermore, the majority had a Masters De gree, had five or less years of youth therapy experience, and identified cognitive/behavioral to be their primary th eoretical orientation. While the demographic characteristics of the present sample are relatively consistent with reported norms for the mental health workfo rce (SAMHSA, 2002), the sample is distinct in several respects. First, it is noteworthy that therapists highest educational degrees ranged from high school diploma (or General Equivalency Diploma) to Masters-level, but none of the participants had doctoral or medical degrees. S econd, while therapists youth therapy experience ranged from less than one year to more than 15 years, over 50%
25 of the sample had five or less years. Lastly, it is notable that no therapists in the present study identified their primary theoretical orie ntations to be psychodynamic/analytic or humanistic, and only 12% of the sample identifi ed their primary theoretical orientation to be eclectic. Table 1. Therapist Demographic and Background Information Therapist Variable N % Total 42 Age (in years) Missing <30 30-39 40-49 60+ 2 10 16 13 1 4.8 23.7 38.1 31 2.4 Sex Male Female 12 30 29 71
26 Table 1. (Continued). Therapist Variable N % Race/Ethnicity White/Caucasian Black/African American Asian/Pacific Islander Other (Multiracial) 33 7 1 1 78.6 16.7 2.4 2.4 Parent Status (children?) Yes No 25 17 59.5 40.5 Highest Degree Completed High school diploma/GED Associates Degree Bachelors Degree Masters Degree 1 1 12 28 2.4 2.4 28.6 66.7
27 Table 1. (Continued). Therapist Variable N % Youth Therapy Experience Less than 1 year 1-5 years 6-10 years 11-15 years Over 15 years 2 21 7 9 3 4.8 50 16.7 21.4 7.1 Primary Theoretical Orientation Cognitive/Behavioral Family Systems Eclectic 30 7 5 71.4 16.7 11.9 As shown in Table 2, most of the youths in cluded in the client sample were male and white/Caucasian, respectively. While the clients ages ranged from 6.3 to 17.9 years old, the majority of youths were at least 13 ye ars old. Most of the youths were receiving both individual and family th erapy. Clients presented a ra nge of clinical problems at intake. Based on the therapists reports, 43% of the client sample demonstrated more than one type of presenting problems, and al most 30% presented with a combination of internalizing and externalizing symptoms. Wh ile the co-occurrence of internalizing and externalizing disorders is considered to be quite common, reported prevalence rates vary
28 widely across studies (Oland & Shaw, 2005), and the rate found in the present sample falls within the expected range. Table 2. Client Demographic and Background Information Client Variable N % Total 42 Age (in years) Missing 6-9 10-13 14-17 4 4 11 23 9.5 9.5 26 55 Sex Male Female 24 18 57 43 Race/Ethnicity White/Caucasian American Black/African American Hispanic/Latino Other (Multiracial) 31 4 2 5 73.8 9.5 4.8 11.9
29 Table 2. (Continued). Client Variable N % Clinical Symptoms Internalizing Externalizing Developmental Substance Abuse Other 18 32 1 6 8 43 76 2.4 14.3 19 Treatment Modality Missing Individual only Family only Individual + family 4 9 5 24 9.5 21.4 11.9 57.1 Measures Therapist Background Form (See Appendix B). Therapists provided demographic information (i.e., age, sex, race /ethnicity, parent status) and professional information (i.e., highest degree completed, years of youth therapy experience, primary theoretical orientation) about themselves on an 8-item, pencil-and-paper survey comprised of both multiple-choice and fill-in-the-blank questions.
30 Case Information Form (See Appendix C). Therapists provided demographic information (i.e., age, sex, r ace/ethnicity) and clinical information (i.e., types of presenting problems) about their respective clients, and specified the treatment setting (i.e., IFP, TFC, RTF/TGH, OPC) and modality (i.e., individual, family, both) for each case, on a 9-item, pencil-and-paper survey comprised of both multiple-choice and fill-inthe-blank questions. Maslach Burnout Inventory Levels of professional burnout were measured with the Maslach Burnout Inventory Human Se rvices Survey (MBI; Maslach & Jackson, 1981). This 22-item, paper-and-pencil questionnaire asks therapists to indicate how frequently they experience specific job-relate d feelings, using a 7-point Likert-type scale (0=never; 6=everyday). Ratings are used to calculate subscale sc ores representing the three dimensions of burnout: Emotional E xhaustion (EE), Depersonalization (DP), and Personal Accomplishment (PA). While some items are associated with more than one dimension, scores for each subscale are c onsidered to be independent and are not combined into a single total score. The EE subscale is comprised of 13 items and yields a potential score range of 0 to 78. The DP s ubscale is comprised of 17 items and yields a potential score range of 0 to 102. The PA s ubscale is comprised of 14 items and yields a potential score range of 0 to 84. A highe r degree of burnout is represented by higher scores on the EE and DP subscales, but lower scores on the PA subscale. Numerical cutoffs (see Table 3) may be used to further classify individuals subscale scores as representative of low, moderate or high degrees of bur nout, based on a normative sample of professionals in human service fields (i.e ., education, social services, medicine, mental
31 health, other). This measure has been wi dely used and has demonstrated extensive empirical support. The MBI manual (3 rd edition) reports Cronbachs alphas of 0.90 for EE, 0.79 for DP, and 0.71 for PA (Maslach, J ackson & Leiter, 1996). In the present study, reliability coefficients were 0.91 for EE, 0.69 for DP, and 0.75 for PA. The correlations between the three MBI subscales for the present sample and the normative sample are relatively co nsistent (see Table 4). Table 3. MBI Subscale Score Classifications EE DP PA Low 0-16 0-6 39-84 Moderate 17-26 7-12 32-38 High 27-78 13-102 0-31 Table 4. MBI Subscale Correlation Matrix DP PA EE Study sample Normative sample 0.548** 0.520** -0.135 -0.220* DP Study sample Normative sample -0.330* -0.260* Note: p .05, one-tailed; ** p .01, one-tailed
32 Counselor Self-Efficacy Scale-Modified Version (See Appendix D. A modified version of the Counselor Self-Efficacy S cale (CSES; Melchert, Hays, Wiljanen, & Kolocek, 1996) was used to measure therapis ts perceived self-efficacy for counselingrelated activities. This pencil-and-paper survey asks therapists to indicate their levels of agreement with 20 statements, using a 5-point Likert-type scale (1 =strongly disagree; 5=strongly agree). The last five items of the original CSES were removed, due to lack of relevance for the present study (i.e., self-e fficacy for conducting group interventions), and replaced with five novel items designed to tap into therapists self-efficacy for counseling activities not addressed by the firs t fifteen items (e.g., I am not adequately prepared to bridge cultural differences during the counseling process). Total scores for this measure were computed by reverse coding negatively phrased items and then calculating the sum across all items. As suc h, higher total scores represent higher levels of perceived self-efficacy for counseling-related activities. This measure has a minimum score of 20 and a maximum score of 100. No norms are currently available for the CSES; however, research suggests that the original measure has high internal consistency and test-retest coefficients, as well as strong construct validity (Melchert, Hays, Wiljanen, & Kolocek, 1996). The 20-item modified version of the scale administered in the present study yielded an acceptable Cronbachs coefficient alpha of 0.88. In order to further evaluate the psychometric properties of the modified measure, the analyses were re-run excluding three of the five novel CS ES items, which had item-total correlations below 0.40. This cut-off was chosen based on statistical standards described by Spector (1992). Cronbachs coefficient alpha for the 17-item scale slightly increased to 0.89. In
33 addition, the analyses were run including only the fifteen items from the original CSES, which yielded a slightly lower Cronbachs co efficient alpha of 0.86. A comparison of the correlation matrices generated for the 20-item, 17-item, and 15-item scales, respectively, revealed that the relationships between CSE and the other study va riables (i.e., EE, DP, PA, masculinity, femininity, therapeutic alli ance) did not differ si gnificantly depending on which CSES version was used. Given that the 17-item version (h ereto referred to as the CSES-M) had the highest internal cons istency, it was used for all subsequent analyses. Bem Sex Role Inventory The original version of the Bem Sex Role Inventory (BSRI; Bem, 1979) was used to measure therap ists levels of masculinity and femininity, and to classify therapists into gender role orientation categor ies. This self-report, paperand-pencil survey asks respondents to ra te a total of 60 items (20 representing stereotypically masculine tra its, 20 representing stereotypica lly feminine traits, and 20 filler items representing neutral traits), using a 7-point, Likert-type scale (1=never or almost never true of me; 7=always or almo st always true of me). Scoring of this measure occurred in multiple stages. First, individuals ratings on the masculine and feminine items, respectively, were averaged in order to create two di stinct raw scores Second, the means were converted into t-scores ( M =50, S.D. =10), representing individuals Femininity standa rd scores and Masculinity st andard scores. Finally, the Difference Score/Median-Split Hybrid Met hod described in the BSRI manual (Bem, 1979) was used to further classify therapists into gender role orie ntation categories. Difference scores were obtained by subtracti ng individuals Masculinity standard scores
34 from their Femininity standard scores. Di fference scores outside the range were considered sex-typed (i.e., gender-typed), with positive difference scores representing a Feminine gender role and negative differen ce scores representing a Masculine gender role. Individuals with difference scores that fell within the 10 range were classified either as androgynous or as undifferentiated, ba sed on a median split. Therapists whose Femininity and Masculinity scores both fell above the mean were defined as androgynous, while all others were defined as undifferentiated. In contrast to the traditional classification method, which categ orizes individuals as stereotypically masculine, stereotypically feminine, or non-se x-typed, the hybrid technique used in the current study allows for further classifi cation of non-sex-typed individuals into androgynous and undifferentiated gender role ca tegories. The BSRI has been used extensively throughout the field and shown to have acceptable internal consistency, testretest reliability, and construct valid ity (Choi & Fuqua, 2003; Holt & Ellis, 1998). Cronbachs coefficient alpha was 0.89 for the present sample. Child Therapy Bond Scale (See Appendices E and F). Primary therapists and clients provided their respective perceptions of the therapeutic alliance on parallel versions of the Child Therapy Bond Scale (CTBS; Shirk & Saiz, 1992). This 7-item, paper-and-pencil questionnaire asks respondent s to rate the qualit y of the therapeutic relationship, using a 4-point Likert-type s cale (1=not like you/your patient; 4=very much like you/your patient). Negatively phras ed items were reverse coded before total scores were computed by summing the ratings across items. Previous research has reported the internal consistencies for th e therapist and youth forms to be 0.79 and 0.85,
35 respectively (Shirk & Saiz, 1992). Based on the present study sample, Cronbachs coefficient alphas were 0.90 for the therapist version and 0.81 for the client version. The therapist and client versions of the CTBS were correlated 0.59 ( p .01), which is higher than values previously reported in th e literature (e.g., K azdin et al., 2006). Procedures The archival data used in the present study was originally collected as part of an ongoing quality management initiative ta king place within the aforementioned organization. Data collection for internal research is standard practice for this organization. As such, participants did not receive financial compensation for their involvement. The present study was carried out in accordance with professional and legal standards of ethical conduct for research involving human subjects. In order to protect the anonymity of participants, therapists a nd clients were assigne d unique numbers for data identification purposes and no additional identifying information (e.g., names, addresses) was provided to th is researcher by the organization. To encourage honest responding and further insure that participan ts privacy was safeguarded, therapists and clients did not have access to one another s responses. This policy extended to the CTBS, on which both therapists and clients evaluated their alliances. The University of South Florida Institutional Review Bo ard provided approval for this study. Upon assignment of a new youth therapy case, therapists were each given a packet containing the following study measur es (in order): Therapist Background Form, Case Information Form, CSES-M, MBI, BSRI, and CTBS-therapist version. Therapists
36 were asked to complete all measures, except for the CTBS, prior to the intake session. Both therapist and client participants were asked to complete the CTBS following the third session in which the part icipating youth client was pres ent for at least 15 minutes. This time point was chosen, as later ratings of the alliance might have been confounded with clients therapeutic im provement, and previous rese arch has shown third-session alliance ratings to be a robus t predictor of youth treatment outcomes (Shirk & Karver, 2003). Participating clients were provided w ith the CTBS-client vers ion by treatment site staff. Furthermore, clients under the age of 11-years-old, as we ll as illiterate and particularly low functioning cl ients, were assisted by sta ff members (other than the primary therapists) at the treatment sites, to assure that the measure was completed properly. To encourage honest responding, therapists and clients were asked to complete the CTBS in separate rooms/locations, and we re reminded that their responses would not be shared with one another. Participants returned their completed measures in sealed envelopes to the clinical dire ctor/C.E.O. of the organization. Identifying information was removed before copies of the data were prov ided to this researcher. It was initially planned for all primary therapists at the orga nization to participate in the study with one or more newly assigned youth clients. Howe ver, due to poor return rate and response errors, only data from 42 ther apist-client pairs (i.e., one cas e per therapist) was included in the present study.
37 Results Descriptive Statistics MBI-Emotional Exhaustion : Scores on the EE subscale (Table 5) are normally distributed, although therapists reported relatively low levels of emotional exhaustion, on average. In fact, 70% of the samples scores fall in th e low burnout range ( 16), while less than 10% fall in the high burnout range ( 27), for this subscale. Furthermore, significant restriction of range is apparent, as the highest to tal score (35) falls well below the maximum possible total score of 78. Th ese findings are inconsistent with higher levels of EE reported in several previous st udies (Ackerley et al ., 1988; Rupert & Baird, 2004). However, a t -test revealed that the score distri bution for the present sample is not significantly different from that reported in the MBI manual for the normative sample of mental health professionals ( t =1.0501, p=0.29). MBI-Depersonalization: Scores on the DP subscal e (see Table 5) are normally distributed, but therapists repor ted relatively low levels of depersonalization, on average. Similar to the EE subscale, almost 70% of the therapists DP scores fall in the low burnout range ( 6), while less than 10% fall in the high burnout range ( 13). Range restriction again is evident in that the highe st total score reported by the present sample (16), falls significantly below the maximum po ssible score of 102. Furthermore, a floor effect was found, as over 14% of the samples total scores equal zero on this subscale. This is inconsistent with pr evious research that found highe r levels of DP (Ackerley et al., 1988; Rupert & Baird, 2004). Yet, a t -test revealed that the score distribution for the
38 present sample is not significantly different from that reported for the normative sample ( t =0.4229, p =0.67). MBI-Personal Accomplishment : Consistent with the other two MBI subscales, PA scores (see Table 5) are normally distributed, although ther apists reported relatively high levels of personal accomplishment, overall. Restriction of range is apparent, as all the therapists PA scores fall between 28 and 48. Furthermore, whereas approximately 67% of the samples scores fall in the low burnout range ( 39), less than 5% fall in the high burnout range ( 30). Once again, this is inconsis tent with lower levels of PA reported in some other studies (Ackerley et al., 1988; Rupert & Ba ird, 2004). Unlike the EE and DP subscales, however, a t -test reveals that therapists in this study had significantly higher PA than that re ported for the normative sample ( t =8.1059, p<.0001). Counselor Self-Efficacy Scale-Modified Version As shown in Table 5, scores on the 17-item CSES-M are normally distributed, but significant range restriction is evident. In fact, while there is a minimum possible tota l score of 17 for this measure, the lowest score found in the present sample is 54. The da ta reflects that therapists in this sample had particularly high levels of perceived se lf-efficacy for counseling related activities. This is not consistent with lower levels of CSE found is some prior samples (e.g., Melchert et al., 1996) Bem Sex Role Inventory : Descriptive data for the Fe mininity (FEM), Masculinity (MASC), and Femininity-Masculinity Difference (F -M) scales are shown in Table 5. It is noteworthy that only 26 therapists BSRI scores were included in the present study, due to missing or invalid ratings (many therapists incorrectly completed this measure on their
39 clients, rather than themselves). T -tests and chi-square test s were used to evaluate whether there were underlying differences betw een therapists whose BSRI ratings were included and those whose BSRI ratings were missing or excluded (n=16). No significant differences between these two groups were found on any of the other measures. The present sample was also compared to the normative sample, as described in the BSRI manual. T -tests indicate that the present sample is not significantly different from the combined-sex normative sample in terms of femininity ( t (840)= 0.7624, p =.44), masculinity ( t (840)=0.2212,p=.82), or femininity-mascu linity difference scores ( t (840)=0.669, p=.50). Nor were significant differences found when male therapists were compared to the normative sample of males (femininity: t (484)=1.61, p=.11; masculinity: t (484)=1.735, p=.08; femininity-masculinity differences: t (484)=0.075, p=.94) and female therapists were compared to the normative sample of females (femininity: t (354)=0.402, p=.69; masculinity: t (354)=1.884, p=.06; femininity-masculinity differences: t (354)=1.709, p=.09). The observed distribution of gender role orientations in this sample was compared to the expected proportions (see Tabl e 6) reported in the BSRI manual for the combined sex normative sa mple, using chi-square tests. The results are not statistically significant ( X 2 =1.02, p=.80). In addition, the observed distributions of gender role orientations fo r males and females in the present sample were compared to the expected proportions reporte d in the BSRI manual for the normative sample of males and females, respectively (see Table 6) Neither the comparison for males ( 2 =2.059, p=0.56), nor the comparison for females ( 2 =3.527, p=0.32), yielded statistically significant results, suggesting that the present sample is relatively consistent with the
40 normative sample. However, it is important to note that significant lack of power likely influenced these findings, as chi-square calc ulations are only considered reliable when the expected value is five or higher, and this assumption was violated in these analyses. Child Therapy Bond Scale: The descriptive data for the therapist and client versions of the CTBS are shown in Table 7. The distribution of scor es on the CTBS-T is normal and ranges from the upper to the lower lim its of the scale. However, there is an over-representation of high scores, sugges ting that most therapists had positive perceptions about the strength of their alliances with clients. Due to poor return rate, only 36 client ratings of the al liance were available. In or der to evaluate whether there were underlying differences between res ponders and non-responders on the CTBS-C, t tests and chi-squares were used to compare th e two groups across the other variables. No significant differences were found. In contrast to the CTBS-T, ther e is less variability represented in the distribution of client scores. There is also evidence of range restriction, as the lowest clie nt rating (13) fell significan tly above the minimum possible score of 7. This suggests that clients tended to report positive percepti ons of the alliance. Related, a ceiling effect was found in that over 5% of the client allianc e ratings fell at the upper limit of the scale. This is consistent with several previous studies that included client ratings of the alliance (e.g., Creed & Kendall, 2005; Kendall, 1994; Kendall, Flannery-Schroeder, Panichelli-Mindel, & Southam-Gerow, 1997; Shelef, Diamond, Diamond, & Liddle, 2005; Shirk & Karver, 2003).
41 Table 5. Descriptive Statistics for I ndependent Variable Measures* MBI-HSS CSES-M BSRI EE DP PA Total FEM MASC F-M ** N 42 42 42 42 26 26 26 Possible Score Range 0-78 0-102 0-84 17-85 1-140 1-140 Minimum 2 0 28 54 67 74 -28 Maximum 35 16 48 84 117 133 35 Mean 15.41 5.41 39.00 73.33 94.61 99.58 -2.00 SD 8.55 4.62 5.37 7.56 13.37 14.09 14.45 Note: Based on raw sum scores; ** F-M= Femininity-Masculinity Difference Score
42 Table 6. Gender Role Orientation Rates Current Sample (N=26) Normative Sample (N=816) n % n % Androgynous Females Males Total 6 0 6 37.5 0 23.1 82 67 149 24.1 14.1 18.3 Undifferentiated Females Males Total 5 4 9 31.3 40 34.6 84 157 241 24.7 33.0 29.5 Masculine Females Males Total 2 4 6 12.5 40 23.1 34 194 228 10.0 40.8 27.9 Feminine Females Males Total 3 2 5 18.8 20 19.2 140 58 198 41.2 12.2 24.3
43 Table 7. Descriptive Statistics for the CTBS Therapist Client N Valid 42 36 Missing 0 6 Possible Score Range 7-28 7-28 Minimum 7.00 13.00 Maximum 28.00 28.00 Mean 19.9524 21.6667 SD 4.47188 4.10575 Hypothesis Testing It was hypothesized that the three MBI subscales (EE, DP, and PA) would be significantly intercorrelated. This hypothesis was supported for all but one comparison. The correlation between EE and PA is not statistically significant (see Table 4). Second, it was hypothesized that EE and DP scores would be negatively correlated with CSES-M scores, whereas PA scores would be positively correlated with CSES-M scores. Pearson correlations were calculated between CSES-M and each of the three burnout subscales (See Tabl e 8). While all of these relationships are in the expected
44 directions, only the correlation between PA and CSES-M scores reaches significance ( r =0.383, p .01, one-tailed). Table 8. Intercorrelations be tween Independent and Dependent Variables CSES-M EE DP PA MASC FEM CTBS-T CSES-M 1 EE .061 1 DP -.102 .548** 1 PA .383** -.135 -.330* 1 MASC -.098 .032 -.127 .291 1 FEM .297 .005 .001 .540** .107 1 CTBS-T -.140 -.204 -.267* .369** -.017 .101 1 CTBS-C -.216 -.225 -.145 .251 -.113 -.012 .590** Note: p .05; ** p .01 Third, it was hypothesized that the dimens ions of professional burnout would be correlated with gender role or ientation and that androgynous th erapists would have lower levels of burnout than stereotypically mascu line and stereotypically feminine therapists would. Table 9 shows the means and standard deviations of the three burnout dimensions for each gender role group. ANOVAs were computed for each burnout dimension, none of which yielded statistically significant results (EE: F (3,22)=0.571, p=.64; DP:
45 F (3,22)=0.716, p=.55; PA: F (3,22)=0.751, p=.53). Femininity and Masculinity scores also were examined as continuous variable s using Pearson correla tions (see Table 8). While masculinity is not significantly correlate d with any of the th ree burnout subscales, a significant positive correlation is found be tween femininity scores and PA scores ( r =0.54, p .01, one-tailed). Table 9. Results for Each Gender Role Orientation Category BSRI Classification N Mean SD Androgynous 6 13.00 6.26 Undifferentiated 9 13.11 9.13 Masculine 6 17.50 7.71 MBI-EE Feminine 5 16.40 5.98 Androgynous 6 4.17 5.67 Undifferentiated 9 5.00 3.61 Masculine 6 7.17 6.34 MBI-DP Feminine 5 8.00 5.43
46 Table 9 (Continued). BSRI Classification N Mean SD Androgynous 6 39.50 6.72 Undifferentiated 9 39.33 5.57 Masculine 6 35.50 3.02 MBI-PA Feminine 5 37.00 6.78 Androgynous 6 71.33 5.65 Undifferentiated 9 78.11 5.73 Masculine 6 66.33 8.80 CSES-M Feminine 5 73.40 5.27 Androgynous 6 20.83 4.49 Undifferentiated 9 16.78 3.90 Masculine 6 19.83 6.74 CTBS-T Feminine 5 20.00 4.47 Androgynous 5 21.00 5.70 Undifferentiated 7 18.43 1.99 Masculine 5 23.40 4.39 CTBS-C Feminine 5 21.80 4.76
47 Fourth, it was hypothesized that CSES-M sc ores would be correlated with gender role orientation and androgynous therapists would have higher levels of CSES-M, relative to therapists with masculine or feminine gender roles (see Table 8). An ANOVA was calculated, yielding statis tically significant results ( F (3,22)=4.134, p =.02). Post hoc analyses (Least Square Difference tests) reve aled that androgynous th erapists scores are not significantly different from the other gender role groups on this measure, as hypothesized. Rather, a statistically significant difference was found between undifferentiated and masculine therapists, with the former group reporting slightly lower counseling self-efficacy. It is possible that this is a chance finding, due to the large number of tests run. Femininity and masculin ity scores on the BSRI also were examined as continuous variables using Pearson correlations (See Table 8). Femininity scores demonstrate a moderate positive correlation with CSES-M scores, although this relationship falls below levels of statistical significance ( r =.297, p=0.07). On the other hand, masculinity scores ar e not significantly related to CSES-M scores. Fifth, it was hypothesized that both therapist and client ratings of the alliance would be negatively correlated with EE and DP scores, but positively correlated with PA scores. Pearson correlations were calculated separately for therapist and client alliance ratings (see Table 8). Therapist ratings on the CTBS were found to be significantly correlated with DP (r =-0.267, p<.05, one-tailed) and PA ( r =0.369, p<.01, one-tailed) scores, but not with EE scores. Client ratings on the CTBS ar e not significantly correlated with any of the MBI subscales. It is noteworthy, however, that EE and PA scores are moderately associated with client ratings of the alliance, in the expected
48 directions, but the correlations fail to r each levels of statistical significance ( r =-0.225, p=0.09, one-tailed and r =.251, p=0.07, one-tailed, respectively). Sixth, it was hypothesized that therapist and client ratings of the alliance would be positively correlated with CSES-M scores. Pearson correlations were calculated separately for therapist and client alliance ratings (see Table 8). Neither of these correlations approaches levels of statistical significance. Finally, it was hypothesized that therapist a nd client ratings of the alliance would be related to gender role or ientation and androgynous therapists would have stronger alliances than masculine or feminine ther apists would (see Table 9). Separate ANOVAs were calculated for therapist and client alliance ratings, neither of which yielded statistically significant results (CTBS-T: F (3,22)=1.022, p=.40; CTBS-C: F (3,18)=1.457, p=.26). When femininity and masculinity we re examined as continuous variables in relation to therapist a nd client alliance ratings, signific ant relationships were not found (see Table 8). Because gender role orient ation failed to demonstrate significant relationships with the other variables, the proposed mediation analyses were not performed. Post Hoc Analyses Having identified DP and PA to be signi ficantly correlated with the CTBS-T, a standard multiple regression was conducted to determine the amount of variance in alliance ratings accounted for by these dimens ions of burnout. The overall regression equation was significant ( F (2,39)=3.716, p .05; R 2 =.16). Upon examining the partial
49 correlations, however, it was determined that DP did not uniquely account for a significant amount of variance in therapist allia nce ratings ( r 2 =.03, t =-1.047, p=.30). On the other hand, the unique contributions of PA were found to be st atistically significant ( r 2 =.09, t =2.031, p<.05). In order to further evaluate the data therapist demographic variables were examined in relation to the variables of interest using one-way ANOVAs and t -tests with Bonferroni adjustments to reduce family-wise errors. Differences by therapist sex were significant for one variable. Specifically, highe r levels of femininity were reported by female therapists than male therapists ( t (24)=-2.933, p=.007). This finding is consistent with prior research on the BSRI (e.g., Bem, 1981). It is noteworthy that the present study did not support previous findings that burnout tends to be higher in male therapists than in female therapists, particularly on the DP subscale (Maslach & Jackson, 1985). In order to evaluate differences by age, therapis ts were divided into three age groups (<30, 30-39, 40). No differences were found. In addition, no differences were found by race/ethnicity, parent status, or years of youth therapy expe rience. As for theoretical orientation, therapists who reported having a family systems orientation had significantly lower masculinity scores on the BSRI in co mparison to cognitive-behavioral therapists ( p=.009; Bonferroni corrected alpha=0.017). Analyses to examine differences across treatment settings were not performed due to the small number of participants drawn from each of the non-home-based programs.
50 Discussion The current study aimed to fill a gap in th e literature by evaluating the empirical relationships between several pretreatment therapist characteristics (i.e., professional burnout, counseling self-efficacy, and gender ro le orientation) and the therapeutic alliance in youth mental health treatment. It was hypothesized that therapist and client ratings of the therapeutic alliance would demonstrate negative relationships with emotional exhaustion (EE) and depersonalizat ion (DP), but positive relationships with personal accomplishment (PA) and counse ling self-efficacy (CSE). Androgynous therapists (i.e., those with high levels of masculinity and fe mininity) were expected to have lower burnout, higher CSE, and stronger al liances, in comparison to therapists with masculine or feminine gender role orientations, irrespective of therapist sex. The results of this study provide support for several of thes e hypotheses. Most notably, significant relationships were detected between therapist ratings of the alliance and two of the three burnout domains. A small to medium effect size was found for DP, as therapists who reported hi gher levels of depersonalizati on (i.e., a bias towards making negative, impersonal, and dehum anizing attributions about cl ients) tended to rate the alliance less positively. A medium effect size was found for PA, as therapists who reported higher levels of personal accomplis hment (i.e., positive feelings and attitudes about ones professional abilities and achieve ments) tended to rate the alliance more positively. Furthermore, PA appears to be a uni que predictor of therapist alliance ratings. These findings may underestimate the true stre ngth of the relationships between these dimensions of burnout and therapist allian ce ratings, given that significant range
51 restriction is apparent on these measures. Also, it is not eworthy that a small to moderate positive effect (r=.251, p =.07) was found between therapis t ratings of PA and client ratings of the alliance, although the correlation was not found to be statistically reliable. It is possible that with greater power a significant effect ma y have been detected. Taken together, these results suggest therapists views a bout their clients and themselves make an important contribution to the quality of alliances formed in youth treatment. While causality cannot be inferred from the present results, these findings underscore the need for greater considera tion of therapists work-related feelings, attitudes, and perceptions. It is possible that depersonalization and most specifically diminished personal accomplishment interfere w ith therapists abili ties to develop strong alliances with their clients. That is, ther apists who experience these aspects of burnout may be less likely to exhibit traits and behavi ors that tend to foster the alliance and/or more likely to exhibit traits and behaviors that tend to hinder the alliance. For instance, research suggests that explori ng clients subjective experien ces is positively associated with the alliance (Karver et al., manuscript unde r review). It is possible that therapists who make negative, impersonal, and dehumanizi ng attributions about their clients (i.e., demonstrate elevated DP) may be less likely to attend and respond to clients individual feelings and perceptions, thereby jeopardizing alliance development. In addition, research has shown that pus hing clients is negatively a ssociated with the alliance (Creed & Kendall, 2005). Perh aps therapists who view thei r professional achievements to be inadequate and unfulfilling (i.e., de monstrate diminished PA) experience cognitive dissonance, which motivates them to seek validation of their therapeutic abilities by
52 increasing their clients clinical improvement In turn, such therapists may become overly vigilant in their efforts to elicit change in their clie nts, at the expense of building rapport and nurturing the alliance. Research examining burnout in re lation to specific therapist behaviors (measure d through observational coding methods) and the alliance is needed in order to identify the most critical behavioral manifest ations of burnout. By extension, it is possible that a dynamic relationship exists between burnout and therapists abilities to form positive allian ces with clients. That is, therapists who have previously experienced difficulties deve loping alliances with clients, or had poor success rates with respect to client outcomes not necessarily due to the therapists own levels of competence, may develop negative perceptions of themselves and/or their clients (i.e., signs of burnout). If such nega tive attitudes and feelings do in fact interfere with therapists effectiveness, they may be le ss able to form positive alliances with future clients. Continued failure to develop str ong relationships with clients may reinforce therapists critical views of themselves and th eir clients, and thus th eir levels of burnout may be maintained or increased. Longitudinal studies are needed in order to investigate whether this sort of circular and self-perpetuating relationship exists. Preliminary evidence suggests that prior expe riences with clients do play a role in the development of burnout. For instance, res earch has shown that therapists who sense inequity or a lack of reciprocity with thei r clients experience a decrease in perceived levels of personal accomplishment (Truchot et al., 2000). Related, Bakker et al. (2006) found that high neuroticism and low extr aversion predicted higher levels of depersonalization for volunteer counselors w ho reported many negativ e experiences with
53 clients, but not for those who reported few negative experiences with clients. Similarly, high neuroticism and low extraversion, respec tively, predicted lower levels of personal accomplishment for volunteer counselors who reported many negative experiences with clients, but not for those w ho reported few negative experiences with clients. A limitation of that study, however, was the use of retrospective self -reports to measure therapists prior experiences with clients, rather than prospective methods and multiple informant ratings. By examining therapists characteristics and experiences over the course of their professional training and careers, it may be possible to isolate variables that predict and/or vary w ith levels of burnout. Increas ed understanding about how internal and external variables interact over time to produce symptoms of burnout may allow researchers to identify risk and prot ective factors that co uld be targeted in prevention or intervention efforts. One of the three burnout domains, emo tional exhaustion, was found not to be significantly related to the therapeutic alliance, in th e present study. The correlations between emotional exhaustion and both therapis t and client ratings of the alliance are in the negative direction, as expected, but ar e not statistically reliable, (CTBS-T: r =-0.20; p=.10; CTBS-C: r =-0.23; p=.09). These findings sugge st that, compared to the other dimensions of burnout, emotional exhaustion is less related to the quality of alliances formed with youths. It is possible that cert ain factors serve to pr otect the alliance from the effects of emotional exhaustion in ther apists. For instance, perhaps therapists naturally, or through clinical tr aining and experience, are able to separate their personal feelings from their work with clients. As such, therapists who experience mild to
54 moderate emotional exhaustion, as did most therapists in the present sample, may retain the ability to contain their distress during sessions and not let it interfere with their interactions with clients. Alternatively, it is possible that therapists do show emotional exhaustion during sessions, but that these symp toms do not have a consistently negative impact on clients. Therapists who experience emotional exhaustion may be more able to understand and sympathize with their clients di fficulties in order to facilitate alliance development. No research examining this theory was found in the existing literature. Another possibility is that th erapists who self-disclose or otherwise evidence signs of mild personal distress to thei r clients may strengthen the alliance by both validating and challenging clients. This notion is supporte d by Linehan (1993), who describes therapist self-disclosure to be an important part of conducting dialectical behavior therapy (DBT) with clients with borderline personality disorder. In cer tain circumstances, Linehan encourages DBT therapists to share their feelings of frustration with clients. For instance, she suggests a therapist might sa y: When you call me at home and then criticize all of my efforts to help you, I feel frus tratedI start thinking you dont really want me to help you (377). Related, some c lients may view therapists as more genuine and relatable if they self-disclose or otherw ise evidence mild signs of personal distress. The empirical research in this area, albeit limited, has provided some empirical support for therapist self-disclosure in general. For instance, Hill a nd colleagues (1988) found therapist self-disclosure was associated with clients positive eval uations of therapist helpfulness. Knox and colleagues (1997) found th erapist self-disclosure to be associated with clients insight and per ceptions of the therapist as mo re real and human, which in
55 turn were associated with the therapeutic relationship. Barrett and Berman (2001) found that clients liked their therapists more a nd had less distress associated with symptoms following treatment, when their therapists e ngaged in self-disclosure in response to similar client self-disclosure. Despite th e positive findings reported in some studies, other studies have reported negative or neutral relations hips between therapist selfdisclosure and therapeutic outcomes (Hill & Knox, 2002). More research examining the relationship between EE and therapist behaviors, such as self-disclosure, is needed before conclusions may be made about the ro le of EE in alliance development. When interpreting the present findings re garding EE, however, it is important to consider that the small size of the present sample afforded insufficient power to detect small to medium effects. In addition, lack of variability associated with range restriction on this scale may have prevented relationships with the alliance from being detected. This would help to explain the nonsignificant correlation found between EE and PA, a finding that is not consistent with previous research (M aslach & Jackson, 1981). Only six therapists in the present sample were cl assified with high burnout on this subscale, and all of their EE scores fall relatively cl ose to the cut-off (and well below the maximum possible score). That therapists in the pres ent sample tended to have low to moderate levels of EE, is consistent with some previous studies (e.g., Rosenburg & Pace, 2006), however other studies have found highe r levels (e.g., Maslach & Jackson, 1981 ) Interestingly, Rupert and Morgan (2005) found that professional psychologists (i.e., clinicians with PhDs) are at greatest risk for emotional exhaustion, of the three burnout
56 dimensions. The present study suggests th at these findings may not generalize to therapists without doctoral degrees. Another explanation for the low levels of EE reported by the current sample is that therapists with hi gher levels of burnout may have been less likely to participate in the research (i.e., they may not have returned th e measures) if they viewed the task as an additional stressor. Related, it is possible that no therapists with high levels of EE were employed at the time of the present study, as previous research has shown this burnout dimension to be associated with turnover rates in agency settings as well as intentions to leave the mental health field (Brown a nd Pranger 1992; Lloyd & King, 2004; Rosenberg & Pace, 2006; Soderfeldt et al. 1995). The c linical director/C.E.O of the organization that provided the current sample noted that turnover rates within the organization are lower than rates typically found in community -based settings, but those therapists who resign typically cite financial pressures, wo rk-family conflict, or unspecified personal problems as their motivations (Reay, personal communication, August, 2006). Furthermore, he indicated that most therapis ts who have resigned reported intentions to leave the field altogether or enter private practice in order to increase their incomes and their abilities to work flexible hours. While the validity of these reports cannot be confirmed, it is noteworthy that therapists explanations for terminating employment at the organization are consistent with signs of emotional exhaustion. Another explanation for the low levels of EE reported by the current sample relates to environmental factors. Research has shown that work-set ting characteristics are significantly related to levels of stress and burnout (e .g., Rosenberg & Pace, 2006; Rupert
57 & Morgan, 2006). It is possibl e that the organization from which the current sample was drawn provides a positive and supportive work a tmosphere that buffers against the effects of professional stressors, such as working with difficult clients. This would be consistent with Butler and Constantines (2005) findings that school-ba sed counselors with higher collective self-esteem (that is, more positive perceptions of their colleagues and the supports available in their work environments) reported lower levels of emotional exhaustion. The clinical director/C.E.O. indicated that he and other supervisors actively work to prevent and alleviate burnout in thei r trainees and less expe rienced therapists by promoting an atmosphere of supportiveness a nd reciprocal communica tion. Further, he reported that stress reduction and collegial ity among therapists, supervisors, and other administrators is encouraged through sponsorsh ip of social activities outside the work environment (e.g., team bike rides). Given that prior research has shown em otional exhaustion to have potentially serious implications (e.g., Edelwich & Brodsky, 1980; Rosenberg & Pace, 2006), it is important for more studies to investigate this variable. As with the other dimensions of burnout, an important next step will be to examine the rela tionships between emotional exhaustion and therapists in -session behavior. Knowing more about how emotionally exhausted therapists actually behave will he lp promote better understanding of how this dimension of burnout functions with resp ect to the therapeutic process. Research has demonstrated that burnout can have extensive physical, emotional, interpersonal, and attitudinal implications for professiona ls (Kahill, 1988), as well as serious financial and bureaucratic repercus sions for organizations faced with staff
58 turnover and shortages (Evans et al., 2006). Moreover, it has been suggested that allowing therapists with signi ficant symptoms of burnout to continue practicing presents ethical concerns, as the quality of services provided to their clients may decline (Enochs, 2004; Rupert & Morgan, 2005). It is not surprising, therefor e, that various methods of addressing burnout have been s uggested in the literature. Some authors have emphasized ways that individual therapists can reduce their symptoms of burnout or their risks of deve loping burnout in the future. For example, therapists are encouraged to set boundaries on their therapeu tic responsibility and resist tendencies to take ownership of their cl ients problems (Friedman, 1985; Kaslow & Shulman, 1987). Related, it has been suggested that therapists work to establish balance between their professional i nvolvement and their personal liv es. Engaging in exercise (Freudenberger, 1974), maintaining healthy eating habits (Raquepaw & Miller, 1989), taking regular vacations (Mas lach, 1976), participating in personal psychotherapy (Fleischer & Wissler, 1985; Kaslow & S hulman, 1987; Piercy & Wetchler, 1987), and developing strong networks of social s upport (Maslach, 1978; Patterson, Williams, Grauf-Grounds, & Chamow, 1998) have all be en recommended as potential ways to manage work-related stress that can lead to burnout. Suggestions have also been made for prevention and intervention strategies at the organizational level. For instance, Ma rtin and Schinke (1998) recommend that orientation programs and in-service training workshops be used to address issues of professional burnout. The authors also sugge st that supervisors and administrators promote an atmosphere of open communication and exchange of constructive feedback.
59 Other suggestions include limiting the time therapists are required to spend on administrative tasks (Raquepaw & Miller, 1989 ), decreasing work hours, and otherwise decreasing workload (Pines & Maslach, 1978). Unfortunately, these recommendations may be unrealistic given the current financial and political pressures organizations face. Perhaps more practical are Selvini and Selvini-Palazzolis (1991) suggestions that employers encourage collaboration, team cons ultation, and emotional connection within the workplace as ways to buffer against stre ssors that can lead to burnout. Professional supports may be particularly important for tr ainees and new therapists. In a study of burnout in marriage and family therapists, Rosenberg and Pace (2006) found that more seasoned clinicians reported less burnout and le ss use of supports. The authors note: It is possible that these professional supports do, in fact, buffer agai nst the effects of burnout, but that the effectiveness of that buffe r is greater for less-e xperienced clinicians or those newer to the field (96) More research in this area is needed in order to test the relative effectiveness of the many recommende d strategies for pr eventing and reducing burnout at different settings and points in therapists careers. Given the conceptual overlap and signif icant correlation f ound between CSES-M scores and PA scores (a dimension of bur nout), it is particularly surprising that counseling self-efficacy did not demonstrate ev en a positive trend wi th respect to the alliance. While items comprising the PA scal e of the MBI ask about feelings related to professional accomplishment (e.g., I feel exhi larated after working closely with my recipients), items on the CSES-M ask about specific competencies (e.g., My knowledge of behavior change principles is not adequate). Accordingly, therapists self-
60 appraisals of their knowledge and skills may be less important than their attitudes about their professional roles and their senses of fulfillment from their work. These results suggest that therapists per ceptions of their own counselin g-related skills may not be linked to their actual abilities to form relationships with clients. It is possible that some therapists with positive views of themselves convey over-confidence to their clients, who in turn perceive their therapists to be condes cending or insincere. It is also possible that therapists with high CSE tend to engage in less self-monitoring and/ or tend to be less attentive to their clients responses to treatment. It is important to note, however, that a number of factors may have affected the present findings. For instance, significant range restriction on the CSES-M may have provided insufficient variability to allow fo r a relationship with the alliance to be detected. Given that no therapists in the pr esent sample rated themselves to have low counseling self-efficacy, it remains unclear whet her very low levels of CSE might harm the therapeutic alliance with youth clients. The same environmental variables that may account for low levels of burnout in this sample may also help to explain the relatively high levels of CSE. More specifically, the cl inical director/C.E.O. noted that he and the other supervisors attempt to c onvey their trust in therapists by encouraging them to take risks and think outside the box with respect to clients. In addition, the organization reportedly works to foster therapists self -confidence and levels of comfort in their therapeutic roles using Socratic methods of supervision. Therapists educational background also may have contributed to their hi gh CSE. Most of the therapists in this sample had recently earned degrees and were accumulating supervised clinical hours
61 towards licensure. It is possible that r eaching this critical professional milestone increases therapists senses of achievement and competence. More research is needed in order to determine whethe r the present finding of no relationship between CSE and the alliance is found in other samples. It would be beneficial for future studies to utilize larger samples that include therapists with a wider range of CSE levels. Furthermore, it will be important for future studies to examine whether CSE is more important for alliance fo rmation when working with certain client populations. For instance, research has indicated that children with externalizing behavior are often difficult to engage in treatment (He nggeler, Schoenwald, Borduin, Rowland, & Cunningham, 1998) and less likely to develop positive alliances with their therapists (Fields, Handelsman, Karver, & Bi ckman, manuscript in preparation). Farber (1990) suggests that non-mutuality in the ther apeutic relationship can evoke feelings of dissatisfaction and frustration in therapists Similarly, Truchot et al. (2000) found a decrease in perceived levels of competence and self-efficacy when therapists perceived inequity or a lack of reciprocity with their clients. Accordingly, it is possible that high pretreatment CSE may be more critical when working with clients who have externalizing problems, as therapists confid ence in their skills may help buffer against development of apathy or insecurity about thei r abilities to reach such clients. Related, it is important for future studies to examine whether ratings of CSE ar e more critical when they are inconsistent with therapists actual counseling abilities. It is possible that therapists who demonstrate strong therapeutic skills, but view their abilities as inadequate (i.e., have low CSE), may have worse allian ces than those who accurately rate their
62 abilities to be high. Similarly, therapists w ho perceive themselves as highly efficacious, but demonstrate poor counseli ng abilities, may have wors e alliances than those who accurately rate their abilities to be poor. This could be tested by using independent observer and client ratin gs of therapists behavior to measure counseling abilities. If inconsistencies between CSE a nd performance are shown to be important for alliance formation, it will be important to examine whether providing therapists with additional supervision and training may assist th em in balancing these factors. The hypotheses that androgynous therapists would have lower levels of burnout, higher levels of CSE, and stronger alliances were not supported in the present study. These findings suggest that ha ving high levels of both masculine and feminine traits may not be implicated in therapists professional functioning. This is incongruent with prior studies showing androgyny to be associated with various in dices of adaptive functioning, including strong interpersonal sk ills. Perhaps gender roles function differently within the unique context of therapy. For instance, it is possible that traini ng and experience allow therapists to demonstrate both masculine and fe minine traits in thei r professional roles, regardless of their self-p erceived gender role orientation. It is noteworthy that femininity demonstrates a significant positive relationship with PA in the current study. One possible explanation for this fi nding is that therapists with higher levels of femininity experience greater fulfillment and have more positive attitudes because their professional roles are more aligned with stereotypically feminine ideals (e.g., developing strong relationships, being emotionally supportive, etc. ). Interestingly, while masculinity is not significantly related to any of the other variables in this study, therapists with masculine
63 gender role orientations reporte d higher counseling self-efficacy than did therapists with undifferentiated gender role orientations Upon closer inspection, however, the difference between the groups means is rela tively small and may not be clinically meaningful. While no significant differences were found between the present sample and the normative sample on the BSRI, it is possible that the sample size was not large enough to be representative and reliable. T hus, more research examining the distribution of gender role orienta tions in therapists is needed, as it remains unclear whether such professionals differ from the gene ral population on this trait. None of the hypothesized relationships w ith client ratings of the alliance are supported in the present invest igation. These findings suggest that youths generally did not view their alliances less favorably when their therapists reported elevated symptoms of burnout and/or low CSE. It is note worthy, however, that EE and PA scores demonstrate small to moderate relationships with CTBS-C scores, but these correlations fail to reach levels of statis tical significance. Several fa ctors may help explain these findings. First, the analyses between the therapist variables and the CTBS-C involved cross-informant comparisons, which often yi eld lower correlations relative to singleinformant comparisons (Karver et al., 2006). Second, lack of variability and range restriction on the CTBS-C and other measures limited the statistical power for detecting relationships. Third, there were a substant ial number of cases for which CTBS-C data was not available, thus further limiting the already modest power for detecting relationships with therapist characteristics. These factors may help account for the fact that therapist ratings, but not client ratings, of the alliance were found to be correlated
64 with DP and PA. While no significant diffe rences between CTBS-C responders and nonresponders were found across the other variables, it remains possible that those clients who did not provide responses on the CTBS-C were less engaged in treatment and had less positive perceptions of the alliance. A dditional research is needed in order to determine if these findings are replicated in larger samples. While the alliance has been shown to pred ict treatment outcomes, it is important that further research examines the direct re lationships between therapist characteristics and youth treatment outcomes. The child and adolescent literature continues to lag behind the adult field with re spect to identifying common process factors, such as therapist traits, that account for variability in treatment outcomes. The therapist variables considered in the present study have been almost entirely neglected in the youth treatment field. Yet, these variables may ha ve unique implications for working with child and adolescent clients. If burnout counseling self-efficacy, and gender role orientation are found to be associated with youth treatment outcomes, it will be interesting to examine whether these relati onships are mediated by other common therapy process factors such as the therapeutic alliance. In addition to the aforem entioned caveats associated w ith the small sample size, missing/invalid data on the BSRI and CTBS-C, range restriction on all measures except the BSRI, and the correlational nature of th e findings, the present study has several other limitations that warrant mention. First, use of self-report measures to evaluate therapist characteristics and, perhaps more importantl y, the alliance introduces the potential for response bias (e.g., social desirability bias) and confounding explanations for results. For
65 instance, clients may have provided inflated alliance ratings if they believed that this was expected of them by their parents. Therapis ts reports of their sy mptoms of burnout and CSE, as well as their alliance ratings, may reflect response bias if they were motivated to appear well-adjusted and professionally competen t. Related, it is possible that therapists who tend to have a biased response style are more likely to rate their perceptions of themselves, their work experiences, and their re lationships with clients in a consistently positive or negative manner. Research has shown that halo effects are potential confounds when therapist charac teristics and the alliance are assessed by the same person (Ackerman & Hilsenroth, 2003). Using behavi or observation to rate the alliance would remove the bias and other problems asso ciated with self-report measures. Another limitation of the current study is that alliance ratings were available for only one case per therapist. Although sampling errors s hould be equally distributed across the therapists, without multiple alliance ratings per therapist, it was not possible to examine within-therapist variance. By eval uating the relationships between pretreatment therapist characteristics and the alliance in mu ltiple concurrent cases per therapist, it may be possible to determine whether burnout, CSE, and gender role orientation are more important when treating clients with particular demographic or clinical characteristics. Other limitations in this study relate to the use of archival data. The data used in the present investigation was collected with in a relatively small organization that provides services in various settings, including clients homes. As such, it was not possible for data collection procedures to be closely supervised. The organization was not able to monitor if and wh en each participant completed the measures. It is possible
66 that therapists did not complete the MBI, CSES-M, and BSRI prior to intake and that therapists and clients did not complete the CTBS after their third face-to-face therapy sessions together. A number of participants did not return thei r completed measures immediately following the third therapy session, as instructed, but returned them later. Although participants were asked to date each measure, few provided this information. As such, it was not possible to verify whether all of the meas ures were completed at the appropriate times. Moreover, some of the therapist and client measures were never accounted for. Such data collection issu es, which are common when conducting realworld applied research, are problematic as they introduce potentially confounding explanations for results. In partnering w ith community agencies to carry out studies, researchers do not always have control over the implementation of research procedures, highlighting the need for better education a bout the importance of following standardized research procedures for staff and service providers in community-based settings. Finally, the fact that the present samp le was drawn from a single organization introduces the possibility that these findings may not generalize to other settings. Therefore, it will be important for future st udies to attempt to replicate the present findings. In addition, further re search is needed in order to investigate whether these findings generalize to therapis ts with more traditional th eoretical orientations, as therapists in the present study were predom inantly cognitive-behaviorally oriented and therefore not typical of most co mmunity-based therapists. It is also noteworthy that most of the therapists were students or recent graduates working towards licensure, and had five or less years of youth therapy experience. Related, therapists education ranged from
67 less than college to graduate level (i.e., Maste rs degree), but none of the participants had doctoral or medical degrees. In these respects the sample reflects a current trend within the mental health industry to hire front-lin e service providers without advanced degrees and with less experience (Ivey, Scheffler, & Za zzali, 1998). Nonethel ess, the findings in this study may not generalize to therapists with more experience and/or higher degrees. Despite these limitations, th e results of the present study provide a jumping off point for further research examining the relationships between therapist characteristics and the therapeutic alliance in youth treatment. Substant ial research has demonstrated that the therapeutic alliance plays a critical role in th e treatment process and has a significant impact on clinical outcomes for clie nts of all ages (Lambert & Barley, 2002; Safran & Muran, 2000; Shirk & Karver, 2003). At this point, however, relatively few empirical studies, particularly in the child and adolescen t field, have investigated potential links between pretreat ment variables and the therapeutic alliance. Identifying factors that are associated with the quality of relationships fo rmed with youth clients is an important first step towards better unde rstanding the mechanisms underlying alliance formation. Furthermore, increasing awareness a bout therapist variables that are related to the alliance may facilitate efforts to improve clinical training and develop more effective interventions for children and adolescents.
68 References Ackerley, G., Burnell, J., Holder, D., & Kurdek, L. (1988). Burnout among licensed psychologists. Professional Psychology Research & Practice, 19 (6), 624-631. Ackerman, S. J., & Hilsenroth, M. J. (2003). A review of therapist characteristics and techniques positively impacting the therapeutic alliance. Clinical Psychology Review, 23, 1-33. American Psychological Association Research Office (2003, January). Demographic Shifts in Psychology. Retrieved September 16, 2003, from http://research.apa.org/gen1.html Antill, J. K., & Russell, G. (1980). A preliminary comparison between two forms of the Bem Sex-Role Inventory. Bakker, A., Van Der Zee, K., Lewig, K., & Do llard, M. (2006). The relationship between the big five personality factors and burnout: A study among volunteer counselors. Journal of Social Psychology 146(1), 31-50. Bandura, A. (1986). Self-effi cacy. In A. Bandura (Ed.), Social Foundations of Thought and Action: A social cognitive theory (pp. 390-453). Englewood Cliffs, NJ: Prentice Hall. Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist, 44, 1175-1184.
69 Barrett, M. S., & Berman, J. S. (2001). Is ps ychotherapy more effective when therapists disclose information about themselves? Journal of Consulting and Clinical Psychology, 33, 3-10. Bem, S. (1981). Bem Sex-Role Inventory Manual. Redwood City, CA: Mind Garden. Beutler, L., Malik, M., Alimohamed, S., Harw ood, T., Talebi, H., Noble, S., & Wong, E. (2004). Therapist variables. In M. Lambert (Ed.), Bergin and Garfields Handbook of Psychotherapy and Behavior Change (5 th ed., pp.227-306). New York: John Wiley & Sons. Boice, R., & Myers, P. (1987). Which setting is healthier and happier academe or private practice? Professional Psychology Research & Practice, 18 (5), 526-529. Bonett, R. (1994). Marital status and sex: Impact on career self-efficacy. Journal of Counseling & Development, 73 (2), 187-190. Bordin, E. (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy, 16 252-260. Byer, C., Shainberg, L., Galliano, G. (1999). Dimensions of Human Sexuality (5 th ed.). Boston: McGraw-Hill Companies, Inc. Choi, N, & Fuqua, D. L. (2003). The struct ure of the Bem Sex Role Inventory: A summary report of 23 validation studies. Educational and Psychological Measurement, 63(5), 872-887. Corey, M., & Corey, G. (1998). Becoming a helper (3rd ed.) Brooks/Cole Publishing Co.
70 Creed, T., & Kendall, P. (2005). Therapist alli ance-building behavior within a cognitivebehavioral treatment for anxiety in youth. Journal of Consulting and Clinical Psychology, 73(3), 498-505. Cushway, D., & Tyler, P. (1996). Stress in clinical psychologists. International Journal of Social Psychiatry, 42 (2), 141-149. DiGiuseppe, R., Linscott, J., & Jilton, R. ( 1996). Developing the ther apeutic alliance in child-adolescent psychotherapy. Applied and Preventive Psychology, 5 85-100. Doucette, A., Boley, L., Rauktis, M. B., & Pleczkowski, M. L. (2004, March). Connecting with youth: Building a ther apeutic relationship Examining the contributions of youth and teacher/counselors. Paper presented at the annual meeting of the Louis de la Parte Florida Mental Health Institute, Tampa, FL. Dupree, P., & Day, H. (1995). Psychotherapist s job satisfaction and job burnout as a function of work setting and per centage of managed care clients. Psychotherapy in Private Practice, 14 (2), 77-93. Edelwich, J., & Brodsky, A. (1990). Burnout: Stages of Disillusionment in the Helping Professions. New York: Human Sciences Press. Enochs, W., & Etzbach, C. (2004). Impaired Student Counselors: Ethical and Legal Considerations for the Family. Family Journal: Counseling and Therapy for Couples and Families 12 (4), 396-400. Enzmann, D., Schaufeli, W., Janssen, P., & Rozeman, A. (1998). Dimensionality and validity of the Burnout Measure. Journal of Occupational and Organizational Psychology, 71(4), 331-351.
71 Evans, R., Mellor-Clark, J., Barkham, M ., & Mothersole, G. (2006). Developing the resources and management support for routine evaluation in counselling and psychological therapy service provision: Reflections on a decade of CORE development. European Journal of Psychotherapy, Counselling and Health 8(2), 141-161. Farber, B. A. (1983). Stress and Burnout in the Human Service Professions New York: Pergamon Press. Farber, B. A. (1990). Burnout in psychothe rapists: Incidence, types, and trends. Psychotherapy in Private Practice, 8 (1), 35-44. Fields, S., Handelsman, J., Karver, M., & Bi ckman, L. (manuscript in preparation). Predictors of therapeutic alliances formed with youth clients and their parents. Finnoy, O. (2000). Job satisfaction and stre ss symptoms among personnel in child psychiatry in Norway. Nordic Journal of Psychiatry, 54 (6), 397-403. Fleischer, J., & Wissler, A. (1985). The th erapist as patient: Special problems and considerations. Psychotherapy: Theory, Research, Practice, Training 22(3), 587594. Fong, M., & Borders, L. (1985). Effects of se x role orientation a nd gender on counseling skills training. Journal of Counseling Psychology, 32 (1), 104-110. Freudenberger, H. (1990). Therapists as men and men as therapists. Psychotherapy, 27(3), 340-343.
72 Friedlander, M., Keller, K., Peca-Baker, T., & Olk, M. (1986). Effects of role conflict on counselor trainees self-statements, anxiety level, and performance. Journal of Counseling Psychology, 33(1), 73-77. Friedman, R. (1985). Making family therapy easier for the therapis t: Burnout prevention. Family Process, 24(4), 549-553. Harvey, J., & Hansen, C. (1999). Gender role of male therapists in both professional and personal life. Sex Roles, 41( 1/2), 105-113. Hayes, M. (1984). Counselor sex-role valu es and effects on attitudes toward, and treatment of nontraditional male client s (Doctoral dissertation, Ohio State University). Hellman, I., & Morrison, T. (1987). Practice se tting and type of caseload as factors in psychotherapist stress. Psychotherapy: Theory, Research, Practice, Training 24(3), 427-433. Henggeler, S., Schoenwald, S., Borduin, C ., Rowland, M., & Cunningham, P. (1998). Multisystemic Treatment of Antisocia l Behavior in Children and Adolescents. Guilford Press. Heppner, P., & Gonzales, D. (1987). Men counseling men. In M. Scher, M. Stevens, G. Good, & G. Eichenfield (Eds.), Handbook of Counseling & Psychotherapy with Men (pp. 30-38). Newbury Park: Sage Publications. Hill, C. E., Helms, J. E., Tichenor, V., Spiegel, S. B., OGrady, K. E., & Perry, E. S. (1988). The effects of therapist resp onse modes in brief psychotherapy. Journal of Counseling Psychology, 35, 222-233.
73 Hill, C. E., & Knox, S. (2002). Self-disclosure. In J. C. Norcross (Ed.), Psychotherapy Relationships That Work (pp. 255-265). New York, NY, US: Oxford University Press. Holt, C., & Ellis, J. (1998). Assessing th e current validity of the Bem Sex-Role Inventory. Sex Roles, 39 929-941. Horvath, A., & Bedi, R. (2002). The alliance. Psychotherapy relati onships that work: Therapist contributions and responsiveness to patients (pp. 37-69). Oxford University Press. Ivey, S., Scheffler, R., & Zazzali, J. (1998). Supply dynamics of the mental health workforce: Implications for health policy. Milbank Quarterly 76(1), 25-58. Jome, L., & Tokar, D. (1998). Dimensions of masculinity and major choice traditionality. Journal of Vocational Behavior, 52 120-134. Kadushin, A. (1976). Men in a womans profession. Social Work, 21 (6), 440-447. Kahill, S. (1988). Symptoms of professional burnout: A review of the empirical evidence. Canadian Psychology, 29 (3), 284-297. Karasek, R., Triantis, K., & Chaudhry, S. ( 1982). Coworker and supervisor support as moderators of associations between ta sk characteristics and mental strain. Journal of Occupational Behaviour 3(2), 181-200. Karver, M. S., Handelsman, J. B., Fields, S ., & Bickman, L. (2005). A theoretical model of common process factors in youth and family therapy Mental Health Services Research, 7(1), 35-52.
74 Karver, M. S., Handelsman, J. B., & Fields, S. (2006). Meta-analy sis of therapeutic relationship variables in youth and family therapy: The evidence for different relationship variables in the child and adolescent treatment outcome literature. Clinical Psychology Review, 26 (1), 50-65. Kaslow, F., & Schulman, N. (1987). How to be sane and happy as a family therapist: The reciprocal impact of family therapy teach ing and practice and therapists' personal lives and mental health. Journal of Psychotherapy & the Family 3 (2), 79-96. Kazdin, A., Holland, L., & Crowley, M. (1997) Family experience of barriers to treatment and premature termination from child therapy. Journal of Consulting and Clinical Psychology, 65 453-463. Kendall, P. (1994). Treating a nxiety disorders in children: Results of a randomized clinical trial. Journal of Consulting and Clinical Psychology 62(1), 100-110. Kendall, P., Flannery-Schroeder, E., Panich elli-Mindel, S., & Southam-Gerow, M. (1997). Therapy for youths with anxiety di sorders: A second randomized clinical trial. Journal of Consulting and Clinical Psychology 65(3), 366-380. Kestnbaum, J. (1984). Expectations for therapeutic growth: One factor in burnout. Social Casework 65(6), 374-377. Knox, S., Hess, S., Petersen, D., & Hill, C. E. (1997). A qualitative analysis of client perceptions of the effects of helpful therapist self-disclosure in long-term therapy. Journal of Counseling Psychology, 44 274-283. Kohout, J., & Wicherski, M. (1999). 1997 Doctorate Employment Survey Washington, D.C.: American Psychological Association Research Office.
75 Kravetz, D., & Jones, L. (1981). Andr ogyny as a standard of mental health. American Journal of Orthopsychiatry, 51(3), 502-509. Lambert, M. J. (2004). Bergin and Garfields Handbook of Psychotherapy and Change. New York: John Wiley & Sons, Inc. Lambert, M. J., & Barley, D. E. (2002) Research summary on the therapeutic relationship and psychotherapy outco me. In J. C. Norcross (Ed), Psychotherapy relationships that work: Therapist cont ributions and responsiveness to patients (pp. 17-32). London: Oxford University Press. Larson, L., & Daniels, J. (1998). Review of the counseling self-efficacy literature. The Counseling Psychologist, 26 (2), 179-218. Larson, L., Suzuki, L., Gillespie, K., Potenza, M., Bechtel, M., & Toulouse, A. (1992). Development and validation of the Counseling Self-Estimate Inventory. Journal of Counseling Psychology, 39 (1), 105-120. Linehan, M. M. (1993). Cognitive-Behavioral Treatment of Borderline Personality Disorder. New York, NY: Guildford Press, Inc Martin, U., & Schinke, S. (1998) Organizational and individu al factors influencing job satisfaction and burnout of mental health workers. Social Work in Health Care 28(2), 51-62. Maslach, C., & Jackson, S. E. (1996). Masl ach Burnout InventoryHuman Services Survey (MBIHSS). Maslach, C., Jackson, S., & Leiter, M. (1996). Maslach Burnout Inventory Manual, 3 rd edition Palo Alto, CA: CPP, Inc.
76 Matsui, T. (1994). Mechanisms underlying sex differences in career self-efficacy expectations of university students. Journal of Vocational Behavior, 45, 177-184. Meissner, W. W. (1996). The Therapeutic Alliance New Haven: Yale University Press. Melchert, T., Hays, V., Wiljanen, L., & Kolocek, A. (1996). Testing models of counselor development with a measure of counseling self-efficacy. Journal of Counseling and Development, 74 640-644, Munson, W., Stadulis, R., & Munson, D. (1986). Enhancing competence and selfefficacy of potential therapeutic recr eators in decision making counseling. Therapeutic Recreation Journal, 20, 85-93. Nevill, D. (1977). Sex roles and personality correlates. Human Relations, 30, 751-759. Onyett, S., Pillinger, T., & Muijen, M. (1997). Job sa tisfaction and burnout among members of community mental health teams. Journal of Mental Health, 6 (1), 5566. Patterson, J., Williams, L., GraufGrounds, C., & Chamow, L. (1998). Essential Skills in Family Therapy: From the First Interview to Termination Guilford Press. Piercy, F., & Wetchler, J. (1987). Family ^work interfaces of psychotherapists. Journal of Psychotherapy & the Family 3(2), 17-32. Pines, A., & Maslach, C. (1978). Characteristi cs of staff burnout in mental health settings. Hospital & Community Psychiatry 29(4), 233-237. Pleck, J. (1995). The gender role strain paradigm: An update. In R. Levant, & W. Pollack, A New Psychology of Men. New York: Basic Books (11-32).
77 Radeke, J., & Mahoney, M. (2000). Comparing the personal lives of psychotherapists and research psychologists. Professional Psychology Research & Practice, 31 (1), 82-84. Raquepaw, J., & Miller, R. S. (1989). Psychot herapist burnout: A componential analysis. Professional Psychology Research & Practice, 20 (1), 32-36. Rosenberg, T., & Pace, M. (2006). Burnout am ong mental health professionals: Special considerations for the marriage and family therapist. Journal of Marital & Family Therapy 32(1), 87-99. Rupert, P. A., & Baird, K. A. (2004). Mana ged care and the independent practice of psychology. Professional Psychology: Re search and Practice, 35 185-193. Rupert, P., & Morgan, D. (2005). Work setting and burnout among professional psychologists. Professional Psychology: Research and Practice 36 (5), 544-550. Safran, J. D., & Muran, J. C. (2000). Negotiating the Therapeutic Alliance: A Relational Treatment Guide New York: The Guilford Press. SAMHSA Mental Health, Unite d States, 2002 online SMA01-3537) Scher, M., Stevens, M., Good, G., & Eichenfield, G. (1987). Handbook of Counseling & Psychotherapy with Men Newbury Park: Sage Publications. Scher, M. (2001). Male therapist, male clie nt: Reflections on critical dynamic. In G. Brooks & G. Good (Eds.), The New Handbook of Psychotherapy and Counseling with Men (Vol. 2, pp. 719-734). San Francisco, CA: Jossey-Bass.
78 Selvini, M., & Selvini Palazzoli, M. (1991). Team consultation: An indispensable tool for the progress of knowledge: Ways of fostering and promoting its creative potential. Journal of Family Therapy, 13(1), 31-52. Shelef, K., Diamond, G., Diamond, G., & Liddl e, H. (2005). Adolescent and parent alliance and treatment outcome in multidimensional family therapy. Journal of Consulting and Clinical Psychology 73(4), 689-698. Shirk, S., & Karver, M. (2003). Prediction of treatment outcome from relationship variables in child and adolescent therapy: A meta-analytic review. Journal of Consulting & Clinical Psychology, 71(3), 452-464. Shirk, S. & Karver, M. (in press). Assessi ng the alliance in ad olescent treatment: A construct in search of a measure. In D. Castro-Blanco (Ed.), Treatment Engagement with Adolescents. Washington, D.C.: American Psychological Association Press. Spector, P. E., (1992). Summated Rati ng Scale Construction: An Introduction (Quantitative Applications in the Soci al Sciences). Newbury Park, CA: Sage. Steinberg, L. (1996). Adolescence (4 th ed.). Boston, Massachusetts: McGraw-Hill, Inc. Truchot, D., Keirsebilck, L., & Meyer, S. (2000). Communal orient ation may not buffer burnout. Psychological Reports 86(3), 872-878. Vredenburgh, L., Carlozzi, A., & Stein, L. (1999). Burnout in couns eling psychologists: Type of practice setting a nd pertinent demographics. Counselling Psychology Quarterly, 12 (3), 293-302.
79 Wester, S., & Vogel, D. (2002). Working with the masculine mystique: Male gender role conflict, counseling self-efficacy, and th e training of male psychologists. Professional Psychology: Re search and Practice, 33 (4), 370-376. Wester, S., Vogel, D., & Archer, J. (2004). Male restricted emotionality and counseling supervision. Journal of Counseling & Development, 82 91-98. Wisch, A. & Mahalik, J. (1999). Male therapists clinical bias: Influe nce of client gender roles and therapist ge nder role conflict. Journal of Counseling Psychology, 46 51-60. Zohar, D. (1997). Predicting burnout with a hassle-based measure of role demands. Journal of Organizational Behavior 18(2), 101-115.
81 Appendix A Summary of A Priori Hypotheses Hypothesis 1: Professional Burnout and Counseling Self-Efficacy (CSE) a. Emotional Exhaustion (EE) will be negatively correlated with CSE. b. Depersonalization (DP) will be ne gatively correlated with CSE. c. Personal Accomplishment (PA) will be positively correlated with CSE. Hypothesis 2: Professional Burnout and Gender Role Orientation a. EE will be significantly lower for androgynous therapists than for stereotypically masculine or femini ne therapists, regardless of sex. b. DP will be significantly lower for androgynous therapists than for stereotypically masculine or femini ne therapists, regardless of sex. c. PA will be significantly higher fo r androgynous therapists than for stereotypically masculine or femini ne therapists, regardless of sex. Hypothesis 3: Professional Burnout and the Therapeutic Alliance a. EE will be negatively correlated with the therapeutic alliance. b. DP will be negatively correlated with the therapeutic alliance c. PA will be positively correlated with the therapeutic alliance. Hypothesis 4: Counseling Self-Efficacy and Gender Role Orientation a. CSE will be significantly higher for androgynous therapists than for stereotypically masculine or femini ne therapists, regardless of sex. Hypothesis 5: Counseling Self-Effi cacy and the Therapeutic Alliance a. CSE will be positively correlated with the therapeutic alliance.
82 Appendix A (Continued) Hypothesis 6: Gender-role Orienta tion and the Therapeutic Alliance a. Alliance ratings will be significantly higher with androgynous therapists than with stereotypically masculine or stereotypically feminine therapists, regardless of sex. b. EE will mediate the relationship between gender role orientation and the therapeutic alliance. c. DP will mediate the relationship between gender role orientation and the therapeutic alliance. d. PA will mediate the relationship between gender role orientation and the therapeutic alliance. e. CSE will mediate the relationship between gender role orientation and the therapeutic alliance.
83 Appendix B THERAPIST BACKGROUND FORM DIRECTIONS: Answer each of the questions below by circli ng and/or writing-in the responses that best describe you. Please, do not skip any items and make sure to answer each question completely Therapist identification #: ___________ Todays Date: _____________________ 1) SEX: a. Male b. Female 2) AGE: _______________ 3) RACE/ETHNICITY: (Please specify) ______________________________________ 4) DO YOU HAVE ANY CHILDREN? a. Yes b. No 5) WHAT IS YOUR PRIMARY CLINICAL ORIENTATION? (NOTE: this may not match the approach you are using with this case) a. Cognitive and/or Behavioral b. Psychodynamic/analytic c. Family Systems d. Humanistic e. Eclectic
84 Appendix B (Continued). 6) WHAT EDUCATIONAL DEGREES DO YOU CURRENTLY HOLD: (*INDICATE ALL THAT APPLY) a. High school diploma or General Equivalency Diploma (G.E.D.) b. Associates degree Please specify : _______________________________________________ c. Bachelor of Arts/Sciences degree Major(s): _____________________________________________ Minor(s): _____________________________________________ d. Master of Arts/Sciences degree(s) Please specify: _______________________________________________ e. Doctor of Philosophy degree(s) Please specify : _______________________________________________ ________________________________________________ f. Doctor of Medicine degree(s) Please specify :________________________________________________ ________________________________________________ g. Other degrees Please specify : _______________________________________________ 7) TOTAL YEARS OF EXPERIENCE PROVIDING THERAPY: a. Less than 1 b. 1-5 c. 6-10 d. 11-15 e. 16-20 f. More than 20
85 Appendix B (Continued) 8) YEARS OF EXPERIENCE TREATING CHILDREN AND/OR ADOLESCENTS: a. Less than 1 b. 1-5 c. 6-10 d. 11-15 e. 16-20 f. More than 20
86 Appendix C CASE INFORMATION FORM DIRECTIONS: Answer each of the questions below by circli ng and/or writing-in the responses that best describes this case. Please, do not skip any items and make sure to answer each question completely Therapist Identification #______ Todays Date:_______________ 1. Client case #: __________________ 2. Client intake date (DD/MM/YYYY) : _______________________ 3. Client gender: a. Male b. Female 4. Client date of birth (DD/MM/YYYY) : ______________________ 5. Client race/ethnicity: ______________________ 6. Client SES: __________________________ 7. Treatment setting for this case: a. Long-term hospitalization unit b. Inpatient stabilization unit c. Residential treatment facility d. Day treatment clinic e. Outpatient clinic f. Home-based services g. Other: ( please describe: ____________________________________________)
87 Appendix C (Continued) 8. Type of case: a. Individual therapy b. Family therapy c. Both 9. Reason for client referral/presenting problems: (select all that apply) a. Internalizing Symptoms (e.g., de pression, anxiety, withdrawal) b. Externalizing Symptoms (e.g., defiance, hyperactivity/impulsivity, aggression) c. Developmental Concerns (e.g., autism, Asperger disorder, mental retardation) d. Substance Abuse/Dependence e. Other: ( please describe: _________________________________________)
88 Appendix D THE COUNSELOR SELF-EFFICACY SCALEMODIFIED VERSION Therapist Identification #: ________ Todays Date: ______________ Please rate the below statements based on the following scale: 1 2 3 4 5 Strongly Disagree Mode rately Disagree Neutra l/Uncertain Moderately Agree Strongly Agree _____1) My knowledge of personality developm ent is adequate for counseling effectively. _____2) My knowledge of ethical issues related to counseling is adequate for me to perform professionally. _____3) My knowledge of behavior change principles is not adequate. _____4) I am not able to perform psychological assessment to professional standards. _____5) I am able to recognize the major psychiatric conditions. _____6) My knowledge regarding crisis intervention is not adequate. _____7) I am able to effectively develo p therapeutic relationships with clients. _____8) I can effectively facilitate client self-exploration. _____9) I am not able to accurately identify client affect. _____10) I cannot discriminate between meaningful and irrelevant client data. _____11) I am not able to accurately identify my own emotional reactions to clients. _____12) I am not able to conceptualize client cases to form clinical hypotheses. _____13) I can effectively facilitate appr opriate goal development with clients. _____14) I am not able to apply behavior change skills effectively. _____15) I am able to keep my personal issues from negatively affecting my counseling. _____16) I am able to clear ly articulate my interpretation and confrontation responses for clients to understand. _____17) I am able to use different types of clinical respon ses at the appropriate time.
89 Appendix D (Continued) _____18) I am skilled in enough techniques to confront the various problems with which my clients may present. _____19) I am not adequately prepared to bridge cultural differences during the counseling process. _____20) I am able to avoid imposing my personal values on clients.
90 Appendix E CTBS (Client Version) Client ID: ____________________ Therapist ID: _______________________ Assisting staffs/parents signatu re: ______________________________________ We are going to read some sentences about mee ting with your therapist. After reading the sentence, you decide how much each sentence is like you. Is it: (read each of the four possible answers and point to the appropriate statement). Let's try this example: Sample: I play games with my therapist when we meet together. Would you say that is: 1 2 3 4 Not Like You A Little Like You Mostly Like You Very Much Like You (Check on the child's response, e.g., Why do you think that?) Here are the rest. Remember, there are no right or wrong answers, just how you feel. ___________________________________________________________________________ 1. I look forward to meeting with my therapist. 1 2 3 4 Not Like You A Little Like You Mostly Like You Very Much Like You 2. When I'm with my therapist, I want the session to end quickly. 1 2 3 4 Not Like You A Little Like You Mostly Like You Very Much Like You 3. I like spending time with my therapist. 1 2 3 4 Not Like You A Little Like You Mostly Like You Very Much Like You 4. I like my therapist. 1 2 3 4 Not Like You A Little Like You Mostly Like You Very Much Like You 5. I'd rather do other things than meet with my therapist. 1 2 3 4 Not Like You A Little Like You Mostly Like You Very Much Like You 6. I feel like my therapist is on my side and tries to help me. 1 2 3 4 Not Like You A Little Like You Mostly Like You Very Much Like You 7. I wish my therapist would leave me alone. 1 2 3 4 Not Like You A Little Like You Mostly Like You Very Much Like You
91 Appendix F CTBS (THERAPIST FORM) Client ID:_________________ Therapist ID: __________________ Please rate your clients curre nt presentation in therapy on th e following scale. Circle the number corresponding to your rating for each item. 1. The child looks forward to therapy sessions. 1 2 3 4 Not Like My A Little Like Mostly Like Very Much Like Patient My Patient My Patient My Patient 2. The child expresses positive emotion toward you, the therapist. 1 2 3 4 Not Like My A Little Like Mostly Like Very Much Like Patient My Patient My Patient My Patient 3. The child appears eager to have sessions end. 1 2 3 4 Not Like My A Little Like Mostly Like Very Much Like Patient My Patient My Patient My Patient 4. The child likes spending time with you, the therapist. 1 2 3 4 Not Like My A Little Like Mostly Like Very Much Like Patient My Patient My Patient My Patient 5. The child would rather do other things than come to therapy. 1 2 3 4 Not Like My A Little Like Mostly Like Very Much Like Patient My Patient My Patient My Patient 6. The child considers you to be an ally. 1 2 3 4 Not Like My A Little Like Mostly Like Very Much Like Patient My Patient My Patient My Patient 7. The child wishes you would leave him/her alone. 1 2 3 4 Not Like My A Little Like Mostly Like Very Much Like Patient My Patient My Patient My Patient